The following is an entry for the CIMSEC & Atlantic Council Fiction Contest on Autonomy and Future War. Winners will be announced 7 November.

By Michael Hallett

   Knudsen awkwardly grabbed his glasses, the humidity of the tropical island fogging them immediately when he exited the lower level of their “headquarters” and began climbing the external stairs. The stairs wound around the building to provide the only entrance to her office overlooking their single pier. The boss had “bought” the island from a Pacific nation that, disappointed by the continued absence of meaningful US investment in the islands, had begun offering long term leases to anyone with the available money. Laurence bought the entire set and as a result now “owned” one of the largest Exclusive Economic Zones (EEZs) in the Pacific. “Most people think of those countries as tiny islands – I think of them as vast maritime domains with land based, unsinkable headquarters,” Laurence said, justifying her purchase.

   “This is not going to be a fun brief,” he thought. The OPTEST had failed, and he now had to explain why. “Did a detonator timing on the Unmanned Underwater Vehicles (UUVs) cause the failure? Power loss? Did they not have the speed to keep up? But the fishing boat was only going 8 knots!”

   Pausing at the top of the stairs he stared down at his target, the fishing boat they’d “acquired” earlier that week. He hadn’t asked where the fishermen were. It sat at the pier, innocently floating, giving no indication why it had potentially ruined everything he’d worked for. This was his first failure, but Laurence was not the type, he was certain, who suffered failure gladly, allowing for “fast failing” and thus enhanced learning.

   “No, I’m not gating paid to learn, but produce. And I failed.”

   Suddenly realizing they could see him on the landing from inside and might be watching him now, he shoved his glasses back onto his face and opened the door.

   The chief of staff (COS), Laurence’s second in command who was the only person it seemed that had known her for more than a few months stopped talking when Knudsen opened the door. Both he and Laurence looked at him, not speaking. Even the cats stopped and stared at him.

   “Laurence is the master of the uncomfortable silence,” Knudsen thought. He walked over to the table at the side of the screen, placed his tablet down carefully and brought up the brief. Just as he was about to send it to the screen…

   “Well!” Laurence snapped. “Status update!”

   “The test was unsuccess…” Knudsen said.

   “I know that! Everyone knows that. I want to know why? No, I want to know that the next test will work,” she said.

   “I’ve prepared a brief based on the preliminary data…” Knudsen said, returning to his tablet.

   “I don’t care about the brief! Don’t ‘brief’ me, just answer the question!”

   Flustered, and a little lost with the slides to use as a crutch, Knudsen took a deep breath. “Initial reports indicate…”

   “Darn cat!” Laurence yelled, picking the cat up form the table and throwing it across the room. “I told it not to get on the table!”

   “Continue.”

   Knudsen opened his mouth to repsond.

   “You know what? I don’t care. I’m paying you to care.”

   “You are paying him for results,” the COS said.

   “Yeah, you are paid for results, and I’m not seeing any. When is the next test?.”

   “Tomorrow,” Knudsen said.

   The cat jumped onto the table closest to the screen. Knudsen picked up his tablet so the cat wouldn’t step on it. The cat looked at Knudsen disapprovingly from its perch on the table edge.

   As he looked at Laurence, waiting for her response, she reached into her pocket and pointed what appeared to be a laser pointer at the cat.

   Knudsen heard a soft “poof” sound, and watched the cat fall to the floor. A blood and fur mist settled to the table.

   Knudsen froze. “I’m glad I picked up my tablet,” he thought.

   The other cat jumped onto the table next to Laurence.

   “Whose cat is that?” asked the Laurence.

   “Mine,” said the COS.

   “So I just blew mine to hell! I thought it was yours! I loved my cat!”

   The COS’s cat stood on the edge of the table, reconsidering its decision join the discussion.  Unfortunately for the cat, its pre-jump wiggle proved extremely ill advised. Its jump to the ground did not take it out of the Laurence’s line of sight sufficiently quickly, and it expired like its colleague.

   “There. Much better,” said the CEO. “Now we are both unhappy. That is fair.”

   Slowly Knudsen walked to the door and out of the office. When neither the COS or Laurence called him back he hurried down the stairs, thinking of the pre-test checks he still needed to perform – and glad he wasn’t a cat.

***

   The three personnel onboard the old PANAMAX cargo ship were enduring another day of regular steaming. The Rover walked into the Engineering Control Station (ECS) and nearly jumped out of his skin when the Engineering Officer yelled “Boo!” right behind him.

   “Not funny!” the Rover said, as the Engineering officer laughed. “What are you doing down here anyway?”

   “Sometimes I feel guilty getting paid for riding this thing. I come down here to check on you and make myself feel useful,” he said.

   “Yeah, why are we even on this old tub?” asked the Rover.

   “Because it is an old tub,” the Engineering Officer said. “Not worth fully automating. It is cheaper to keep a few people on here.”

   “But nothing ever happens. It is practically fully autonomous,” said the Rover.

   “Yeah, but not completely, and that gap between “Fully autonomous” and “Almost autonomous” costs money to fill, more money than this ship is worth. We’ll ride it for another year or so before they scrap it.”

   “There are worse jobs, I guess,” the Rover said.

   “Yep,” said the Engineering officer. “I’ll see you later.”

***

   The Fillers, (as Knudsen liked to call them) aquatic autonomous vehicles, ascended from their transit routing hugging the ocean floor, homing in on the particular PANAMAX’s merchant ship’s acoustic signature. This signature, a combination of the propeller and machinery sounds, generates a noise ‘fingerprint’ unique to each ship.

   Like a school of salt water piranha dispersing to attack an unlucky cow from multiple directions, the autonomous aquatic vehicles broke off from their transit formation and each headed for a specific location on the ship’s hull.

   The first drone to reach its terminal position attached itself magnetically over a salt water intakes and began releasing a fluid into the water,. This fluid, which looked like a thick milkshake, was sucked into the intake and flowed unimpeded through the pipes, values and pumps for 20 seconds after release from the drone’s tank – then the coagulation began.

***

   “What is going on!” the Engineer muttered to himself as his phone, running the ship Engineering Operational Control System (EOCS) app, suddenly emitted a series of alarms. He was used to the normal alarms, the sequence of beeps that enabled him to in effect listen to the plant and know what kind of minor hiccup it was experiencing, and whether or not he needed to go down to the ECS to deal with it or could manage from one of the many Ship Control terminals spread throughout the ship or in the simplest cases, his phone.

   “Is the Rover messing with me? Revenge?” he thought. “Hysterical.”

   However, when he heard a slight change in the ambient noise generated by the ship, the overall constellation of noises that constituted ‘normal’ in the environment, all of which he was responsible for, he abandoned the login on his stateroom terminal and moving at his maximum speed, headed for ECS.

   As he dogged down the watertight door behind him, usually left open during everyday steaming, he listened to the voice of the ship control system reciting a litany of causalities.

   “#2 firemain pump offline.

   # 2 evaporator offline, #3 offline, #4 offline.

   #1 fire main pump offline. Loss of firemain pressure…”

   “What the …” he said, grabbing the phone to notify the officer the deck (OOD). “She must be freaking out up there too, with all these alarms.”

   The Rover’s voice on the main spaces circuit broke into the repetition of the causalities by the ECS.

   “Firemain is down,” he said frantically, “And everything, I mean everything is tripping offline.”

   “Shutdown the main engines and generators, we’ll run on emergency battery backup!” the EOOW said, placing the phone back on the hook.

   “Aye aye,” said the Rover.

   The EOOW’s hope, that this was a joke imposed on him, already waning, vanished completely.  Relying on the muscle memory instilled by countless engineering casualty control drills, his hands flew over the console executing emergency shutdown procedures. What he couldn’t do from the ECS the Rover was doing in the Main spaces.

   The emergency lighting switched on as he powered down the ship’s service generators. The nearly 1000 foot ship shuddered.

   “It feels like we ran aground!” the Engineer thought. “But that’s impossible.”

   The Ship’s Master entered the space.

   The Engineer caught the Master’s eye but kept working through his shutdown procedures.

   “Shaft speed indicator – zero. The shaft had stopped completely. But even without power it should be turning. Did we lose lube oil pressure?” he thought.

   “Rover, check for catastrophic main reduction gear lube oil loss!” he said over the ECS circuit.

   Two clicks sounded over the circuit as the Rover quickly keyed a handset, the acknowledgement signal during a casualty.

   “What is going on, Engineer?” the Master asked.

   “I don’t know yet sir, we’ve shut everything down. The first indication was a loss of firemain pressure, but the shaft is stopped too, I mean stopped, no movement.”

   The Engineer stood in front of the control consoles, breathing heavily as they both looked for a clue as to the cause of this cascade of causalities. The silence was oppressive. The ship was cold, dark and quiet, the worst possible set of conditions for an engineer.

   “I’m heading to the bridge to make my reports,” the Master said. “Get me answers.”

  “Aye aye sir,” said Engineer.

   The Rover slipped past the Master and dogged down the door in the ECS behind him.

   “I shut everything down, well everything that had not already tripped offline on its own,” the Rover said. “What is going on?”

   “Nothing good,” said the Engineer, as he picked up the handset again and finally called the Bridge.

   “Status report!” the OOD yelled.

   “Total system shutdown, total loss of main propulsion and auxiliaries,” the Engineer said, forcing himself to report calmly.

   “I know that!” shouted the OOD. “Get me power!”

   “The Master is on his way to the Bridge,” the Engineer said. “We’re working on it.”

   He and Rover stared at each other for a moment in the silence.

   “I’m going to inspect the plant, and see if I can figure anything out. Stay here and I’ll report in to you. Break out a log and write down everything I say.”

   The Rover nodded and began looking for the green hardback logbooks that he know were in a drawer somewhere, but that he’d never actually used. The Engineer grabbed another flashlight, dogged the door down behind him, and headed into the Main Spaces, sick to his stomach.

***

   No reports. Knudsen stared at his phone in disbelief. “This can’t fail too. I won’t get off the island alive,” he thought, as he started to sweat. He was due at Laurence’s office in 15 minutes. He’d received the initial reports that the evolution at the merchant ship had started. He’d spoofed the ship’s navigation system to reduce the range the drones had to travel, but not so much that the ship’s course would indicate any sort of relationship to his location.

   “The communication network is good,” he said, checking yet again that the longer range high flying UAVs he’d deployed the week before were still on station, and reporting all conditions normal. The UAVs enabled him to create a secure communication network.

   “When you work for a hacker, you learn to guard against being hacked,” he thought. The high flying solar powered UAVs were the primary transcieving nodes. Other UAV and surface drones, with intermediary UAVs serving as mid-altitude amplifiers, constituted the lower altitude and surface nodes of the communication network. The underwater drones still had to surface to communicate. However, they were so smart communication, once he’d provided his commander’s intent, was seldom necessary.

   “If the comms are good, why am I not hearing anything?” he asked the screen in his hand.

  He checked the network status again. All green.

   “I don’t have time for this! The evolution started, I know that. Did the mid altitude UAV run out of power early? Or just fail?”

    He’d only sent one, in order to reduce the chances of detection.

   “That was obviously a mistake,” he said.

   He checked his watch again.

   “Don’t panic,” he told himself. “What other measures of effect can I find and use?”

   He walked around the FabLab twice before inspiration struck.

   “AIS! I can check the speed on the satellite feed from the ship’s Automatic Identification System!” Pulling it up on his tablet, his eyes watered with joy when he saw the speed – “1.4”

   “Stopped. That only means one thing – success!”

   Calm now, and with 5 minutes to spare, he looked out at his drone Fleets, arrayed in what he liked to refer to as “Centurions,” 10 by 10 squares of surface, subsurface, air and the ones he was most proud of, multi-domain electromagnetically hardened drones.

   He thought of the holding area as his “zoo” in which the various types of bioemulated drones waited for their next mission. By starting with a blank slate in the design process, made possible by additive manufacturing, he’d created machines unlike any made before, blending the robustness and durability of nanotechnology derived materials with biological design principles. The results were drones with all of the advantages of biological systems, without the disadvantages in terms of strength and of course destructive potential.

   “That is not entirely true,” he had to admit. “They still have the energy weaknesses of machines.” Indeed, that was his biggest challenge, how to keep the machines going. It was of course possible ot create drones with long dwell times, in whatever domain, but he not yet been able to effectively combine a long on station time with the capability to act rapidly that Laurence’s plans all seemed to require.

   “But we are still getting the jobs done,” he said out loud to the unresponsive “animal” army arrayed in front of him.

He selected a set of drones and assigned them to the pre-established test mission using his phone. As they departed the FabLab, Knudsen examined his phone.

   “It’s still strange to look at my phone and see only a few apps,” he thought. All the mobile devices ran only on Laurence’s secure work network, which was much less rich, of course, than the normal internet everyone was accustomed to. Neither his phone nor tablet had Web access. All Dark Web access took place through thin clients. Laurence said it was a cyber security measure. Knudsen suspected it was simply a system for controlling the software – the human software.

   “But I’m getting paid enough and creating the machines I’ve always dreamt of, so I shouldn’t complain. I can live without mobile internet access – though I suspect some people couldn’t.”

    As the drones exited the FabFloor he turned and headed up to the Office for the demonstration.

***

   Knudsen walked quickly around the edge of the cliff along the narrow pathway undulating above the jungle below. (the security chief had explained the undulations were designed to provide firing positions, so that one person could, if necessary, hold off any force that achieved the impossible and landed on the island.) The views of the sunrise were amazing, the colors of the biomass below pulsing as the sun rose above the horizon during the golden hour. Laurence had her own view of the sunset, but no one (except maybe the COS) was ever allowed over there.

   He stepped into Laurence’s main work/reception area. She was engaged in a discussion with the COS. He acknowledged Knudsen’s presence with a brief glance, and since he didn’t tell him to “Get the heck out” Knudsen knew it was ok to stay.

   The flow of Laurence’s conversation remained uninterrupted by Knudsen’s entrance. “The power of the movements as promoted in the media is that they accustom people to doing what they are told. Once they are acclimatized to think progressively in one way, say the environment for example, they can be manipulated to think in other ways. There is then so much noise in the system, and people are so unaccustomed to critical thinking, that it becomes extremely difficult for them to support the formulation and execution of appropriate policies.” Laurence said.

   “And your point is,” the COS said.

   Knudsen had to stifle a gasp. He was still shocked every time COS challenged Laurence. But he did, even after the cat incident.

   “They will applaud me as I become the richest person ever, by any measure (purchase power parity, per capita, gini coeficinet), even richer than Xerxes, the only free man in Persia, by making them poorer. Normally people only get rich by providing value to their customers. The richest get that way by providing value to society. I’m going to extract value, and they will love me more because I’ll clothe my actions in the language of virtue to which they have accustomed themselves. They didn’t give me the credit I deserved when I was ‘playing the game’ and they will now pay the price.”

   “And I’ll get rich too,” smiled the COS.

   “Indeed. Not as rich as me, but pretty rich.” They both laughed.

   Laurence turned her brilliant smile onto him. “Ah, my dear Knudsen, are your wee beasties ready for the demonstration?”

   Her ability to act warm and friendly, as if she never got angry, was unnerving. Especially when he knew she wasn’t actually happy to see him. “She will be after everything goes right with this demo,” he thought. 

   “Yes. They’ll reach initiating positions on schedule,” Knudsen said.

   “And how was the other long range test?” the COS asked.

   “Successful,” said Knudsen. “The ship is fully stopped. The engineering plant has been completely disabled.”

   “Very well,” Laurence said.

   Turning to the CSO she asked, “Is the test platform in position?”

   “When you reaching the viewing area you’ll see it,” the CSO replied.

   “Excellent,” Laurence said, hopping up from her chair.

   The fishing boat, captured in the island’s Exclusive Economic Zone, was now 1500 meters off the end of the pier. Laurence had a “fleet” of two yahcts that the island’s government had inducted into, or rather, designated as its “Maritime Law Enforcement Service”. Previously they had relied on their allies to provide EEZ enforcement, with only limited success. The two yahcts had detected and captured the fishing boat the previous week, enabling this second op test of his more overtly aggressive kinetic drones to take place.

   Knudsen snuck a look at his phone’s chart on which the fishing boat and his drones were indicated. Although he didn’t think it was entirely necessary, he’d cheated a bit, and placed a tracker on the boat to ensure that the drones could find it. They would only switch to homing in on the tracker signal if their organic radars and visual detection systems failed to lock onto the target, so he’d still get good data from the test, but it reduced the chances of catastrophic failure as Laurence watched.

   “If the drones flew and swam on by the boat without destroying it, I’m afraid I’d find myself taking a trip to the jungle below at an uncomfortable velocity,” he said to himself.

   He snuck his phone back into his pocket. It all looked good.

   Laurence glanced over at him. “Don’t you have to do something?” she asked.

   “No ma’am, they are fully autonomous. The AI agents will execute the mission as planned.”

   She smiled with her brilliant smile at him, and even though he knew it was just an automatic response of her highly developed charisma, it still made him very happy to have pleased her.

   He snuck his phone out again. Seeing that all the platforms in the engagement were on track, he hit the “execute” button that had appeared mid screen. Pushing that button constituted the final step in the start sequence. If this went well, on subsequent missions he could schedule the final “Go/no go” decision further back giving the AI engines more scope for mission execution, in effect pushing the decision down to the tactical level. He quietly took a deep breath as he waited for the attack to unfold.

   The first indication something happened was the sudden slowing of the fishing boat. Its stern drifted to the right, and it came to an almost complete stop in the calm seas just off the island about 1000 meters.

   “Not too impressive,” muttered the COS.

   “It will become rapidly more glamorous. Underwater drones just removed the ship’s propellers. They intercepted the boat, followed their way back to the shaft seal by adhering to the hull. They then placed a series of three electromagnetic thermite grenades next to the shaft until the grenades magnetically attached themselves. Seconds after attaching, the rotationally activated detonation occurred and the grenades burned through the shaft. Weakened, it snapped, causing the propellers to sink to the bottom, rendering the ship DIW,” Knudsen explained.

   “DIW?” COS asked.

   “Dead in the water,” Knudsen said, “As you can see.”

   “Let’s be honest,” said Laurence. “I was hoping for a bit more.” Her smile was gone.

   “I wanted to demonstrate the effects of each discrete attack. The show is not over yet, and a genuine attack would be planned for a unified ‘time on top’ or in other words, the attacks would occur simultaneously,” Knudsen said.

   Knudsen observed his observation drone take station over the ship and put on his glasses. After confirming the feed was available from the drone, he said, “A close in feed of the ship is now available. You might find it more exciting.”

   Laurence raised her eyebrwos and pulled her glasses down from where they had rested atop her jet black hair.

   The next phase of the attack began with a small explosion. It destroyed the surface search radar – not a big deal on a fishing ship, but the attack provided the proof of concept of “blinding” the target before the follow on drones with larger explosive payloads arrived.

   Next, the bridge windows blew out and a fire started.

   “Not very big explosions,” the COS said.

   “The explosives are sized to cause the appropriate amount of damage to accomplish the objective – destruction of shipboard systems. Each ounce of explosive reduces the range, speed and maneuverability of the drones, so I use the minimum effective dose,” Knudsen said.

   “Makes sense I guess,” the COS agreed. “And makes them hard to see.”

   The ship began listing to port.

   “How’d you do that?” Laurence asked.

    “Sea skimming drones with shaped charges aimed just below the waterline. The ship is also flooding due to the destruction of the shaft seal,” Knudsen explained.

   Two larger drones, that looked like six legged spiders, landed on the deck of the fishing boat. Gun muzzles emerged from the top of their oblong “bodies” and they began shooting the boat’s pilot house.

   “Those enable us to gain control of a ship without destroying it,” Knudsen said.

   “Looks like they are destroying the boat to me,” Laurence said with a happy laugh.

   “These drones provide us with a huge degree of attack flexibility,” Knudsen said with pride.

   “Quite nice, actually,” Laurence said, once again turning her 1000 watt smile onto Knudsen. “I think that’s enough.”

   “Roger,” Knudsen said, taking out his phone. He pushed the “Pause” button.

   “What happens to it now?” COS asked. “Will it just sink right there?”

   “It better not – I don’t want an oil spill on my front yard!” Laurence said.

   “No worries,” replied Knudsen. “I’ve placed high pressure supplemental buoyancy devices on board – basically giant balloons. They will fill the flooded spaces and restore the necessary reserve buoyancy so we can take the ship under tow and dispose of it later. I also removed the fuel. Residual lube oil and fuel remain in the tanks, but the oilphagic microbes I distributed throughout the ship will take care of that.”

   The ship quickly righted itself – it still listed slightly to port, but was clearly no longer in danger of sinking.

   ‘That should persuade people to pay our fees,” said the COS.

   Laurence smiled. “We’ve bent the cost curve so far in our favor they have no choice but to pay. Insufficiently agile, they can’t touch us, and they know the costs of even attempting to build the capability to do so will never be tolerated by their “citizens”.

   Knudsen headed down, happy the demonstration, this time, was visibly and unambiguously successful.

***

    As Knudsen looked out at the FabFloor he’d designed and with the help of his construction units, built single handedly, he felt a great satisfaction mixed with a growing sadness that he was engaged in an enterprise that increasingly appeared to be if not outright criminal was so close as to be indistinguishable from it.

   “I knew when I took the job that Laurence was a major player on the Dark Web, but I let her chic, and the millions of crypto-currency she gave me, cloud my judgment,” he said to himself. “But I have designed and manufactured some amazing technology, and maybe it won’t be as bad as I fear. If I hadn’t agreed to do it someone else would have.”

   Excited, he looked around the Lab and tried to think of someone he could share his good news with. He could chat with the AI on his phone. However, as a programmer he lacked the necessary ignorant empathy based faith possessed by non-programing, regular consumers required to make those sorts of relationships satisfactory.

   “I need an analog solution,” he thought. “I need a dog.”

Michael T. Hallett works in the International Programs department at the U.S. Naval War College and is the Executive Officer of the U.S. Pacific Fleet Maritime Operations Center (MOC) Reserve Detachment 301. Reach him at www.linkedin.com/in/michaelhallettii or michaelhallettii@mac.com. The ideas presented here are those of the author alone and do not reflect the views of the Department of the Navy or Department of Defense.

Featured Image: SAM 3 Minesweeping unmanned surface vehicle (Saab)

The following is an entry for the CIMSEC & Atlantic Council Fiction Contest on Autonomy and Future War. Winners will be announced 7 November.

By Phil Reiman

      The heavy door closed behind him and he poked on the interrogator’s screen. “Interrogator C45-1 Let’s try this again.” An interrogator that answers questions; funny every time, he thought. “Who can get me smart on this?”
      Specify training need, please.
He thought for a second about how to put his request into something the Interrogator could answer, “I need an individual who can bring me up to speed on the battle systems we’re using in this AO.”
      Training Modules on . . .
“Nope, I need a person. I want a tutor”
The Interrogator paused. Milliseconds. Then displayed a list:
      MSG Hanson AC3 – OPS NCO 2301st AABN, assigned TF NORTHERN STRIKE
      SFC Rodriguez MC2 – AOPS NCO 2301st AABN, assigned TF NORTHERN STRIKE
      SFC Palermo-Hang INF3 – BSYS NCO 304th C2IC, assigned TF NORTHERN STRIKE
“Hanson. Have him contact me. Now.” Major Charles Amir leaned back in the padded chair. The green vinyl creaked loudly under his Arctic mil-suit. Even in this armored, insulated command post, the cold was everywhere. Leading in this unit is going to be . . . odd. As a replacement, he didn’t know the politics or personalities – to his left; his right or under him. Worse, he didn’t know the job.
      He knew combat or that’s what his bio said. That’s how he got the job here. Citation-For Actions Before the Enemy – On 13 April, 2024 Lieutenant Charles J. Amir lead 15 Soldiers of the 23rd Armored Infantry Regiment, Heavy (AIRH) in a rearguard action outside of Taipei . . .
      Master Sergeant Hanson calling
      “Master Sergeant, I need help.”
      Gotcha, Sir. Nice to meet you. A youngish Soldier appeared on the screen; dressed in his artic under gear. Just a bump cap; no applique armor no hardware.
      “Where’s your armor?” Amir changed tone, “Uh, sorry for the circumstances.” He came as a replacement for Major Anthony A. Rodriguez-Soto. It was only name to him, but he was their guy; their leader.
      Sir, current SOP is soft suits and bumps inside BCP. Are we changing that?
      “Yes, but I’ll send the message – Interrogator – task on uniform posture. Remind in one hour.
      Acknowledged. Reminder.
      “I am stepping in here and I am stupid on the kind of things this unit can do. I am a heavy rifle and . . . You look young for a Master Sergeant.”
      Automation Corps stuff, sir. Skills are bills; plus, you know I test well and came in under an accelerated program. Most of our Armored Auto Corp Battalion is younger than me.
      A fleeting thought, maturity crossed his mind, but it fled as Hanson added, I read your milbio, sir. I think you’ll be a good fit based on what I read about Taipei.
      “Stop. Taipei was a nightmare. I . . . wait, that was rude. Sorry. I’m still spinning from the trip and this assignment.
      Message Traffic. Welcome. New Mess . . . he cut it off. Ignore.
      Do you need time? Hanson sounded concerned.
      “No, and there isn’t any. I’ll stim. You know what? I’ll summarize what happened in Taipei and then you can get a sense of what I don’t know. Deal?
      Deal.
Hanson told the Soldier how his section of the 23rd AIR drew the rear guard because they were closest to full strength after the chaos outside of the city. Casualties in this conflict came fast and in bunches. His platoon took up positions as advised by their expert systems and of practical necessity a few hundred meters on the friendly side of a canal bridge. They were in roughly in position when alerts flashed. INCOMING. H-sonic; ORG 034 . . . < 1 minutes.
      Teeth gritting. IMPACT.
However, there were only a few earth-shattering kabooms; mainly just a thousand pops. Spider mines, the worst. Clouds of them began to fall around them. They plinked to the ground; flipped open and became a spreading sea of movement, slowing crawling . . . away. Whether a miscalculation or a planned, the moving minefield was well behind his position.   But a few were close. Moving. Those few moved slowly, mindlessly closer.
      These things will eat an unsupported team alive.
Somewhere behind him, a team member either triggered or shot one. Collectively, the spiders reacted, moving wordlessly over the rubble – toward his unit’s position; scratching over the rubble.
      We have minutes. Circle the wagons. Get out and comeback when these things are neutralized. CANDLELIGHT! All ECM on! Flares – Sparklers – Chaff. The mines stopped. Some lurched after flares and detonated. Others wandered aimless. Most sat in a 1000 meter oval around his team.
      Countermeasures will confuse the mines as long as we’re popping, but – we’re lit up. . The enemy will pound as soon as they can get something pointed in our direction. Worse, between battery life and expendables we could hold off these for only a few minutes. The surprise was gone and it doomed his defense Getting out was key.
I called for fire; breach a lane. The shells arrived and churned the earth. The mines were destroyed in the beaten zone, but they began filling in – not too fast but my TAC showed greater density in our egress path. Something made me think they couldn’t follow a head fake. I didn’t think; I put it into motion. Called in another strike – wider, deeper on a new azimuth – 143. The head fake.
      But, I forgot to tell everyone the plan. I dropped a FF marker and I screamed on the coms – everyone to my marker and from there heading 233. Bring all and any! WARPATH-WARPATH-WARPATH! Immediate withdrawal. As the impacts churned the ground, the mines crawled to block the broad passage along 143. The team was moving – fast. I don’t know if you have much experience with exoskeletal-units; everyone knows the ‘iron’ is strong, but when we move – it’s fast.
      Time for the real move. I called in the second strike, pounding the currently lightly covered perimeter along 233 – ninety degrees off the first strike. The arty scattered or killed the few spiders that remained. We pushed up the second lane.
      We had two breakdowns and we triggered a few mines. The iron is great; you’re armored, powered and connected, but if that stuff goes down it can be very lonely. You go from a god of war to a potted plant in an instant. It can trap you. And the shear force that it takes to break them, can squish the mortal inside like ripe fruit. Organs tear loose. So, I keep my TAC to tracking positions and to red status and that’s it. Too much to track – I have been counseled about that, why don’t you have this or that on your TAC? It’s all happening too fast. Turns out that omniscience is overrated. We got out, carrying our wounded and our broken. And they gave me a medal.
      Tracking, sir. Hanson seemed excited.
      He didn’t mention his career before the Second Chinese Civil War or how that sprang from the Black Money crisis of 2021 when banks sold their own crypto currency and how it pulled trillions of dollars out of legit financial system. He skipped how the evaporation of all their citizens’ savings undercut the legitimacy of the PRC. Then coups, counter-coups or revolutions or whatever left the governments of the world trying to avoid the angry lunges from a self-destructive giant. He skipped that the United States was here to help the Chinese against the Chinese and that he was sure in his bones that the invasion of Taiwan was really only an attempt by the northern Chinese to bring us into the war against the southern Chinese. And it had worked.
      He skipped the dramatic advances in material science that made hard, light armor; thin, long-lasting batteries and mobile power plants possible. And the introduction of troop-level AI. Exoskeletons in first world armies. Ubiquitous automated combatants. The Army’s incessant desire to add ‘A’ to every unit designator – the same way that it added ‘mechanized’ to units in the 1930s. This boy-master-sergeant probably knows all this better than I do.
      For a moment, he thought about how the Army he had joined in 2019 was so different now. Sure, everyone had seen all these changes coming, but since managing large organizations was like driving a car by looking in the rearview mirror. Basically have to crash to know when to turn – and the Army was in every way a stodgy organization. This war changed all that.
      A top down, bottom up reconstruction. New units, ranks, careers. Sure you could still be in the infantry, but that would not let you evade the minimum test scores. The new Army was small and expensive and required far more from every Soldier in training and maturity than any other armed force in history. Every decision was ‘assisted’. You were trained on A-systems from your first day in the same why you learned drill and ceremony or marksmanship. And the wash out rate was high. Soldiering isn’t for everyone, buddy; no levee en mass is going to save you from AI guided rocket artillery. Modern armies don’t play with a deep bench. There is no reserve – there’s those ready to fight and – targets, he grinned to himself.
      That assisted leadership was what he was sure had led to this giant, frozen Inchon. The Bad Chinese were paying handsomely in e-coin for anything that it could get over the trade routes in the newly opened arctic. To cut that off, some colonel’s ginned up a course of action that involved a large raiding base near an arctic river basin. Seems reasonable, I am sure that’s how it went down. It’s easy – I have seen it too many times – sit back and let the AMIS or the ATICS or the JORIC or whatever acronym the leader want to hide behind – divine a solution. Pick and click strategy. L-E-D leadership.
      He had little faith in that. You have to go and see; you have to have been there, he thought. That’s what’s missing when we surrender to our AI advisors. The raiding and blockading had been a success at an earlier base but the Chinese had anticipated such a move and came in force to counter it. Leading to a brutal, icy exchange that killed the aforementioned Major Rodriquez and 25 other Soldiers. But they were a strategic head fake – weren’t they? They were the bullet sponge. Designed to draw the bad guys up and out – make them fight and think they’d won. Cold calculation, but that’s what the AI advised. Once the bad guys felt the threat was eliminated, the bottleneck would surge through and we’d pounce. That time was near.
      POSSIBLE PERIMETER ACTIVITY. VECTOR 033.2 TRUE – ASSESSING
      “Hanson, how much time do we have?”
      I don’t know. Bears and foxes set the hounds off almost every day.
“Real bears?”
      Yeah, polar. Biologicals, they call them. The sentry hounds love them. Just like real hunting dogs. At least the hounds were deployed, he thought. The four-legged, sensor units roamed the frozen wilderness at a range of tens of kilometers and would give him a good cushion if something came their way.
      Sir, we’re ready. We’re in place and quiet. We’ve got the latest and greatest. We’ve got a full depot. We’ve got persistent eyes over, on and under the ice. We’ve got long-range, fast drones that can place decoys or batteries where we need them when we need them. Three hours flight time puts us 300 km north. That’s a big net. We are green on power and coms. And we’ve got Mk II cargo loaders to move all this stuff. Give me a convoy and we’ll stampede them to the barbeque.
      ADDITIONAL POSSIBLE PERIMETER ACTIVITY. VECTOR 033.13 TRUE – ASSESSING
      “Should I be worried?”
      No. But . . . , Hanson paused as he brought up a new system.
      The BAAS suggests doubling that sector, by thinning the river side. I’ll shift the hounds to the contact.
The Major watched the screen, “right, can you tell me about your battle system?”
      Sure – like I said, the latest and greatest q-bit wiz bang. It works on the trope-system. . .
“Trope?”
      A pattern recognition system. Like the electric MacArthur that thought up this operation – go after their supply – through the Arctic. Loosely, the system looks for patterns, or conventions in thought and offers solutions based on its weight of different variations.
“I get that. Take it up a notch.”
      The BAAS – or Battle Auto Assistance System states our current situation falls into a bug war. That’s reasonable. The environment here is harsh. Bad for people. Bots don’t care. So we have lots of robots. Lots. The Battle systems sees this as a Bug War – lots of dumb enemy units – and looks for the MacGuffin . . .
“uh?”
      A fragile component that amounts to an Achilles heel. If the system finds it – maybe a lack of redundancy or a probable sensor confusion the BAAS it will suggest it. It draws from mountains of open source and ‘S’ material.
      This kid was educated, he thought. “Does it just suggest to kill the central controller?”
      Right, that would be the Ontological trope. You know, kill the wizard and his minions die? Decapitation in a technospheric sense.
“So, does this thing do more than point out the obvious? Is this thing strategic or tactical? What’s its scale?”
      Above tactical, not granular. Obvious is up to you. . That’s still up to you One guys creative genius is another guy’s hack, but it generates that ‘commander’s intent’ baloney and publishes a sharp looking Opord that . . . oh.
“No issue. The intent is to complete the mission. I don’t believe in end states either.”
      Ok, sorry. My best advice is to think of it as one of those puzzle games. If you take too long, the game wiggles one of the pieces; points to an answer. Gives you a hint.
“Got it.”
      That GUI is everywhere now, if you look for it, you see it all the time now – banks, care centers you name it.
      “Is it like chess AI?”
      Ugh. In chess were talking pattern recognition and as you sort through the moves your looking for patterns. So is the machine. Once it sorts your moves in to a set of patterns, it follows the probabilities to where you are going. But, Chess is not infinite, there are only so many variations. To trick the AI you just have to break the rules. Decide that your piece is a rook or that go under the board. Can’t see that stuff. GEB
What?
Gödel-Escher-Bach. Can’t code inspiration. right?
“That’s Auto Corps slang, right?”
      Right, I’ll. . .
      ADDITIONAL POSSIBLE PERIMETER ACTIVITY – EN FOR 30+ POS . . .
      BAAS INTERRUPT: EN ATTACK. VECTOR 220 < 4 KM
            OVERRUN
                  ADVISE HARD STAND – CUSTER
                  ADVISE SELF DESTRUCT – HERO’S EXIT
                  ADVISE EVADE – BOULDER RUN
                  ADVISE EXFIL – SNAKE HOLE
      Ugh. That’s the advice?! We bought the head fake, he thought. The rest of the attack was a blur when the Major recalled it later. None of the standing orders felt good, but he jabbed at CUSTER. Positions and equipment came up on the screens. Some blue, but a lot of red all coming from the southwest across the frozen river. Auto turrets were tearing into the approaching targets. Lots of kills. Too many at this distance, too close. We can’t shift. In an instant, he knew, we’re not going to stop them.
      The screen flashed with updates as he scanned. Messages to higher for assistance. But the Soldiers, were simply supporting systems – they had to trust those systems to save them and it wasn’t looking good. How do I lead, in a system that is responding without me? Do I go out and pat the machine on the hood, telling the troops this is the best tech around? Inspirationally polish a bumper? I have to DO something.
The red tide was taking enormous punishment as the missile batteries and cannons swept the approaches. Still coming – right at us. Too close. Just as an unfamiliar and unwelcome feeling of panic began to rise he noticed the green units – cargo handlers. Big cargo handlers.
      “Hanson, can you – we – someone move the Mk IIs?!”
      Can . . . I think.
He could. A quick message and the automated cargo hulks were moving at full speed in a variety of directions away from the outpost. The red hoard chased or at least some did. The cargo handlers got farther than he expected. They were tough.   The distraction gave the blue forces the distance and time to thin the attack. Six minutes and forty seconds after the proximity warning, the red attack had dissipated. It felt like he hadn’t taken a breath in days, when he finally filled his lungs with chilly air. Position was blown. They will be back or use larger stand-off stuff in a few minutes.
      Major Amir’s thoughts moved to the next phase; getting everyone prepared to jump out of here, but he gave himself second to savor the moment, fell for the old misdirection play. . .

Phil Reiman is a husband and a father of two living in Alexandria, VA and working as an attorney for the Air Force’s Legal Operations Agency in its Litigation Division at Joint Base Andrews, MD.

Featured Image: Vikhr reconnaissance-assault unmanned ground vehicle with ABM-BSM 30 weapon turret on BMP-3 chassis at Military-technical forum ARMY-2016 (Wikimedia Commons)

The following is an entry for the CIMSEC & Atlantic Council Fiction Contest on Autonomy and Future War. Winners will be announced 7 November.

By Mark Jacobsen

   They were pushing the blackened wreckage of an F-35 over the edge of the carrier deck as the MCV-22 approached to land. The crew had given David the cramped cockpit jump seat, so he had a clear view of the spectacle. He vaguely remembered seeing something like this in a history class once. He did a quick query in his optics. Saigon. That was it. Even with his feeble grasp of history, he had some sense of how humiliating this moment must be for the Navy and for the United States.
  Time slowed as the tiltrotors transitioned, erasing their forward motion, leaving them suspended above the approach end of the USS Gerald R. Ford. The moment was utterly surreal. He was a data scientist. Data scientists did not land on aircraft carriers in active theaters of war.
   He glanced back at his team. The graduate students were craning to see out the tiny side windows, giddy with excitement, recording everything in their optics. They seemed oblivious to the risks they were taking. His older colleagues, who had left spouses and children behind on just a few hours notice, looked more sober. They could die out here. The administrators at Stanford would be having kittens right about now, especially when parents started calling. The team had left suddenly, in the dead of night, on a government helicopter. There was no time for a debate.
   Rylie had sent him pictures, so he knew right where to look as they descended to land. The five heavily damaged F-35s. The scorched arresting gear and electromagnetic launch system. The dual band radar antenna. For the worst U.S. naval disaster since Pearl Harbor, there was so little to actually see. The planes were a mess, but the damage to the carrier looked almost trivial. There were no towers of black smoke, no exploded ammo magazines ripping apart hulls, no twisted masts lurching sickly into waves flickering with burning oil. Nothing to awaken the horror and rage that the occasion probably called for.
   David had never felt strong patriotic sentiment anyways, only a vague sense of gratitude for the comfortable life he enjoyed in the United States. His one experience with the Department of Defense had burned him badly, and he’d never looked back. His was now a cool, abstract world of linear algebra, of Bayesian statistical computation, of natural language processing and training of neural networks, far removed from the passions of war.
   Until yesterday.
   Suddenly all of that mattered immensely.
   The whole war might depend on it.

   Rear Admiral Rylie Marshall Ellis was waiting for them on the carrier deck. She looked regal in her uniform, which David had hardly seen back in the old days, when she mostly wore civvies. She looked much the same as she had then, just grayer. And more exhausted.
   “David,” she said with a thin-lipped smile. Their handshake felt awkwardly formal. Something more familiar seemed called for, but David knew she was still embarrassed about how things had ended fifteen years ago. Even though it hadn’t been her fault, the tension was palpable. “Thank you for coming. I wasn’t sure you would.”
   “It appears my country needs me,” he said.
   “It does indeed. Thank you. I know this can’t be easy.”
No, he thought. It wasn’t. Just seeing her opened all the old wounds.
   “Why me?” he asked her. “Why not go straight to HiveAI? It was their swarm.”
   “That’s exactly the problem. I need an outsider. Someone I can trust.”
   “Was it really that bad?”
   “Let’s get inside. I’ll show you.”

   They had met on a sweltering afternoon at Camp Roberts, California, standing amidst racks and racks of fixed-wing drones, watching the Naval Postgraduate School’s ARSENL team prepare to set a world record. David had been a graduate student at Stanford at the time, developing machine learning algorithms for swarms. Rylie was a naval surface warfare officer studying at NPS. David could still vividly remember the soaring emotion he’d felt when the extensive preflights were finally complete, and the catapult fired airplane after airplane into the sky. The 50 fixed-wing aircraft whirled against the blue like two columns of seagulls. It was extraordinary.
   That day had set both of them on the paths that would define their careers. For Rylie, it was a thesis that drew the attention of the Chief of Naval Operations. For David, it was his company. Inspired by the work at NPS, he had enrolled in the new Hacking4Defense initiative at Stanford, made a name for himself and his team, and won some funding from Special Operations Command. With Rylie’s help, he had incorporated and secured a contract from DIUx, the DoD’s much-vaunted new Silicon Valley outpost.
   It had been the adventure of a lifetime.
   At least until DoD ruined it all.
   After the bankruptcy he had returned to academia, but swarms remained his first love. Those first prototypes at Camp Roberts gave way to far more powerful aircraft, which utilized deep learning to evolve new tactics and maneuvers at millions of times the speed of human thought. An entire industry was devoted to understanding the algorithms the machines devised. Most of that world was classified, but David followed whatever open source materials he could find. The complexity of modern swarm behaviors far exceeded the cognitive capacity of a human mind, and he was lucky if he could get even a cursory sense for their logic. But during those rare moments when a behavior did come into focus, he was inevitably overwhelmed by its mathematical elegance and sublime beauty.
   Which is why the next twenty minutes were among the most disturbing of his life.

   “It started when one of our submersible drones detected a Chinese sub about 50 kilometers out,” Rylie told their group, gathered around a conference table in a tiny room that had been allocated for the investigation. They were linked into a shared virtual workspace, and a 3D reconstruction was unfurled over the table, synthesized from every sensor available in the battlespace. “Our drone requested further assets. The drone carrier launched an airborne defensive swarm and a pack of submersible hunters.”
   David watched the drones launch on the battle map. The Chinese sub surfaced and launched a swarm of its own. This is how the war had been for the past three years. Machine against machine, a bloodless but very expensive game. The swarms converged in a bewildering tangle. Chinese swarms had evolved very similar behaviors as their American adversaries, and nothing about the battle thus far looked unusual.
  “Now watch,” Rylie said. “This is where things get weird. And then it all goes to hell.”
   The Chinese swarm abruptly went haywire. Drones veered erratically, stalled, spun helplessly into the ocean. Lone vehicles fled the fight entirely, only to be shot down by packs of pursuers. Drones turned on each other, meeting in fiery collisions or pursuing each other in futile dogfights while the U.S. swarm easily picked them off. David couldn’t have devised more idiotic tactics if he’d tried.
   And yet something strange was happening. Amidst the confusion, the U.S. formations were breaking apart. Gaps appeared in their impenetrable mesh. Packs waffled between maneuvers, as if confused, something David hadn’t seen since the dark ages of AI. And then, amidst all this weirdness, the Chinese started to score kills. It happened so fast, and in such strange ways, that David could barely follow.
   An automated cry for help went out. The carrier group launched reinforcements, but the next wave failed as badly as the first. The shipboard defenses put up a valiant fight, but too many Chinese drones had survived, and the fleet couldn’t get them all. Six got through. That was enough to kill two deck crewmen and inflict all the damage David had seen while landing.
  A long silence fell over the room. David felt unfamiliar emotion constricting his throat. That taste of Pearl Harbor, which had eluded him before.
   The display vanished. Admiral Rylie Marshall Ellis looked of them in the eye.
   “Your job,” she said, “is to explain what the hell we just saw.”

   They worked the next 36 hours straight.
   The team was exhausted after the first all-nighter and pleaded for rest, but then news arrived of another disastrous engagement near Manila. Four more dead. Rylie showed them pictures.
   “You’re doing it for them,” she said.
   They brewed more coffee.
   Rylie and her aides brought them whatever they needed. Her most precious delivery was five terabytes of highly proprietary data from HiveAI, including the entire codebase for the swarm. It amounted to roughly five million lines spread over thirty-six repositories. David’s coders sank into it like sharks. Meanwhile, his statisticians began building statistical models to capture patterns in the swarm behavior that they couldn’t detect with the naked eye.
   It was almost magical, watching them work. David had always thought the term data science was a misnomer. Really, it was both science and art. This was the art of it. Messy. Creative. Unstructured. You never knew who would make a serendipitous discovery, or how.
   But at the end of the marathon, they had nothing to show.
   The team was broken and demoralized. David gave them twelve hours off.
   He met with Rylie to discuss progress. “We finished our investigation on the drone carrier,” she told him. “The swarm passed a full diagnostic battery hours before the battle. No sign of electronic warfare, either. The communications mesh was up the entire time. All the real-time checksums passed. The swarm appears to have behaved exactly as it was supposed to.”
   “It’s the same at our end,” David said. “The codebase looks clean.”
   “So the devil’s in the algorithms,” Rylie said with a heavy sigh.
   They sat in sullen silence. A clear malfunction would have made things so much easier.
   “Can you find it?” Rylie asked.
   “I don’t know.”
   A long silence ensued.
   Finally Rylie said, “They should have chosen you.”
   David chuckled bitterly.
   His scrappy little startup had been at the cutting edge of swarm intelligence. And then, less than a year after SOCOM and DIUx had funded his company, a new generation of DoD leadership had killed DIUx and terminated the fast-track contracting exemption. That same year, a concerted lobbying effort by the biggest defense companies had rolled back the DoD’s effort to embrace the startup sector. Almost overnight, David’s company unraveled. Saddled with five employees, a year lease on office space, and a product that had little application outside defense circles, bankruptcy was inevitable. The entrenched players had spun off new subsidiaries like HiveAI to scoop up the contracts that entrants like David had lost.
   And they didn’t know the first thing about swarm intelligence. At least in David’s humble view.
   Of course, it was impossible to know how good the training actually was. Given the complexity of machine learning and the high risk of obscure, non-intuitive modeling errors, David had pushed DoD for more openness and transparency in swarm training. But the DoD’s instinct was to overclassify everything, and the defense companies wanted to protect their intellectual property. The old rules were back in place.
   God knew how these machines had been raised, or what they were thinking when they fought.

Three more days passed without a breakthrough.

   Flag officers had been pinging Rylie for updates every ten or fifteen minutes, but now the torrent was slowing. They were losing faith in her. On the third morning, the Pentagon took over the investigation from USPACFLT and awarded HiveAI $50 million to investigate itself.
   Rylie refused to show weakness, but David could see the strain. He hated himself for letting her down, but he had never worked on a problem this hard in his career.
   “What makes it so hard?” the USPACFLT Commander, Admiral Eric Greene, asked them during a call later that morning. David knew he was under incredible pressure himself, but he genuinely seemed interested in understanding.
   “It’s the evolutionary nature of the algorithms,” David said.
   “Keep in mind, I’m a history major.”
   “I’ll give you an illustration,” David said. “You want to teach a simulated robot to walk. You have two options. First, you can manually write a program to articulate every joint just the right amount in the proper sequence. It will probably take you days of tweaking. If you’re lucky, you’ll end up with something that shambles along like a zombie at the end. The second strategy is evolution. You generate thousands of completely random algorithms, each of which articulates random joints by random amounts in random sequences. There’s absolutely no design. Then you try each one out. You have a fitness function that assigns each algorithm a score based on some criteria, such as how far the robot moves. Most algorithms will be a disaster. The robot will just spasm helplessly. But maybe one or two algorithms will show a little forward motion and earn higher scores. Now you create a second generation. You create more random algorithms, but you also keep the high-performing algorithms from the first generation, breeding them and mutating them.”
   “I think I see,” the Admiral said. “The fitness scores increase every generation.”
   “Exactly,” Rylie said. “With enough genetic diversity, a good fitness function, and enough computational brute force, you can breed algorithms that vastly outperform anything a human could design.”
   “That’s why we use them in swarms,” David added. “It’s hard enough for a human to compute how to gain the advantage in a dogfight with a single aircraft. When two swarms collide, a human couldn’t possibly calculate optimum tactics or maneuvers. The only way to develop effective tactics is to evolve them. But there’s a problem. We know they work, but we have no idea why.”
   “What’s worse,” Rylie said, “Is that the evolved algorithms can sometimes be deeply counterintuitive. That’s why we have entire career fields devoted to studying and understanding them.”
   Admiral Greene frowned. “If I’m understanding you, the entire learning process is only as good as your fitness function, right? How can you possibly assign meaningful fitness scores to such complex swarm behaviors?”
   Rylie said, “It’s a design choice. Someone designs the fitness function.”
   “Well, who gets to decide that?”
   David started to speak, paused, and then closed his mouth.
   That was it. That was the answer.
   He was so stupid for not seeing it earlier.
   “I have to go,” he said, rising quickly, leaving the two bewildered admirals behind.

   He was digging through the files from HiveAI when Rylie caught up with him.
   “What did you figure out?”
   “Let me ask you this. How does the DoD design anything?”
   He thought back to his startup all those years ago, trying to do business with the Pentagon. His company had been lean, fast, and agile, and the fast-track contract from DIUx had seduced him into thinking the Pentagon worked the same way. Once the next administration snuffed that little experiment out, he’d had a hard dose of reality. He’d learned about the glacial pace of the acquisitions system, the dozens of layers of bureaucracy needed to approve every decision, the endless caveats and requirements imposed by agencies all over government.
   Rylie saw it now. David could tell by her expression.
   “Do you remember Robbie the Robot?” he asked her. “From Mitchell’s book on complexity?”
   “It was an illustration of genetic algorithms,” Rylie said. “Robbie has to travel around a grid cleaning up scattered pop cans, right? He has to do it in the fewest possible moves. Mitchell showed that genetic algorithms outperform human-designed ones.”
   “That’s right. But there was something else. In the winning algorithms, Robbie leaves cans behind as markers. He goes back for them later.”
   “A classic counterintuitive behavior,” Rylie said.
   “So imagine the Pentagon is designing a trash-collecting robot, and the robot keeps leaving trash around. What happens next?”
   Rylie looked at him with horror.

   Fourteen hours later, David and Rylie were knocking on Admiral Greene’s door at USPACFLT headquarters. Once they knew where to look, it hadn’t taken them long to find evidence. They had flitted from meeting to meeting aboard the Ford, even as they spent the precious minutes in between trying to fill in the gaps in the story. The rest of the team worked furiously on simulations that David requested. When Rylie was satisfied they had enough, they caught a series of flights back to Pearl Harbor.
   They started with Robbie the robot. David had slides.
   When they were satisfied the Admiral understood, Rylie slid a virtual document across to him. She said, “This is a Navy after-action report on AVENGER DAWN, the first open demonstration of the new HiveAI counter-air swarms. During the first dogfight, the swarm voluntarily sacrificed several planes to probe enemy algorithms, at a cost to the Navy of about $3 million. We uncovered emails showing that Admiral Garrett was furious. He replaced the program manager a month later, and personally inserted a requirement that the swarms couldn’t voluntarily destroy their own units without human authorization.”
   “Like telling Robbie he can’t leave his trash as markers,” Admiral Greene said.
   “The Navy imposed nine other requirements in this report alone. And then there’s this report, a year later. Word got out that HiveAI was training Navy swarms using completely random tactics and maneuvers. Someone leaked videos to Navy leadership of early-generation algorithm trials, which were predictably awful. It caused an uproar. The Fighter Weapons School hosted a conference to discuss swarm tactics and training. Every agency sent its best and brightest. That conference resulted in an approved list of material that should serve as the basis for future swarm training.”
   “Wait,” Greene said. “They took out the random variation?”
   “Why rely on randomness, when you have the finest tactics and maneuvers that the Navy’s brightest minds can come up with? There’s more. We’ve only been looking for a few hours, but we’ve identified at least forty-six different caveats applied to genetic variation and fitness functions by sixteen different office symbols.”
   Greene was rubbing his temples now and staring at his desk. “So what does this mean?”
   David took over for Rylie. “I asked my team to run some simulations using an open-source swarm toolkit. They evolved swarm algorithms with the types of caveats we found in the paper trail. Now, this was just a simple toy model, so I can’t tell you how close it matches reality, but the results were dramatic. The constrained algorithms achieved fitness scores about twenty percent of what unconstrained algorithms achieved. And the algorithms were remarkably brittle. When the team pitted them against novel algorithms, the constrained algorithms had no idea how to cope.”
   “Twenty percent,” the Admiral echoed. “We’ve been doing this for ten years. How could we not know?”
   Rylie said, “We train against ourselves. Or our allies, who use similar algorithms. Even our adversaries steal our algorithms. If you pit two of these swarms against each other, they fight marvelously. Maybe one swarm performs just a little better than another, the algorithm improves, and we think we’re one step closer to perfection. Meanwhile, we have no idea we are stuck in the foothills of what might be achievable. And besides, what is HiveAI going to do? Start ignoring contract requirements from the DoD?
   “But imagine if this entire time, somebody was training swarms correctly, letting them evolve organically, totally unfettered. I think every engagement in the plinking war was a sham. The Chinese swarms were deliberately mirroring our tactics, taking huge losses so they could learn and improve their own algorithms. The real battles started this week.”
   Greene’s eyes didn’t leave his desk. He was silent for a long time.
   “What do we do?” he finally asked.

“Did we just give birth to Skynet?” David asked Rylie that evening. They were flying back to the Ford in the morning, but had the evening off, and were enjoying a walk along Waikiki beach.
   Rylie smiled at the joke. “Only if they listen.”
   “You don’t think they will?”
   “Absolutely not. You just asked the United States Department of Defense to let go of everything it has ever known, everything it has ever prepared for, all of its knowledge and skill and mastery of war. You want it to trust the fate of our Pacific Fleet to randomness and evolution, instead of human ingenuity. Of course they won’t listen.”
   Two more battles happened that night.
   In the morning the Secretary of Defense announced the suspension of all genetic algorithms in the United States Armed Forces. Unless granted express authorization, all drone operations would resort back to full human control. This was only a stopgap measure, while the DoD worked with HiveAI and other stakeholders to review swarm training and algorithm development processes. Not long after, the Chief of Naval Operations announced the Navy was pulling Weapons School graduates from billets across the force to augment drone carrier crews in the Pacific and develop new tactics.
   “Do we laugh or do we cry?” Rylie asked David at breakfast.
   “I asked that a lot when my company folded,” David said. “Now I just shrug. Time will tell if DoD can evolve.”
   They ate their breakfast and then sipped their coffee, enjoying the view of Pearl Harbor. David could see the first tourists across the harbor making their way up onto the USS Missouri, which had once been the glory of the Navy’s battleship fleet. She still looked majestic, gleaming in the morning sun.
   A flying V of drones appeared on the horizon. A defensive counter-air swarm, returning from its latest patrol, ever vigilant against America’s enemies. They swept in low over the harbor, made a graceful arc out over the harbor, and eased in to flawless landings.
   They were marvelous to watch, David thought. The most elegant and intelligent machines the United States had ever built.

Maj. Mark Jacobsen (USAF) is a C-17 instructor pilot and SAASS graduate. He is currently completing a PhD in Political Science at Stanford University. He is the author of a novel, a previous story about swarms reprinted through CIMSEC, and has other writings available at buildingpeace.net.

Featured Image: Deck handling trials of the X-47B aboard the USS Harry S. Truman (CVN-75) in December 2012. (Northrop Grumman)

The following is an entry for the CIMSEC & Atlantic Council Fiction Contest on Autonomy and Future War. Winners will be announced 7 November.

By Mark Sable

   Alyssa Wexler always wanted to be captain of her own ship. She relished the idea of standing on a bridge, keeping a cool head in a crisis as the men and women beneath her looked to her for guidance.

   It was Star Trek: Voyager that had stoked her ambition. When she was a child, the sci-fi show envisioned a future where a woman could command a starship.  By the time Wexler was at the Naval Academy, there were no warp drives, but there were women Admirals.

   What Voyager didn’t imagine was autonomy. That when Wexler finally had a command, her crew would be a series of genderless unmanned systems. Her role was more IT manager than captain. The massive supertanker she guided around the Horn of Africa was nothing more than a ghost ship.

   It wasn’t until the automated sonar array picked up a signature it identify that Captain Wexler felt something resembling a purpose. But the anomalous signature was all-too familiar to her. And for the first time in recent memory, Wexler felt something else. Fear.

   The last time she’d felt that sensation so deeply was over 1500 miles to the Northeast, and more years in the past than she cared to admit. Then-Lieutenant Junior Grade Wexler was assigned to an escort vessel in a carrier group deployed to the Persian Gulf to oppose Iranian intervention in the Saudi Civil War. She’d never forget the sight of the Revolutionary Guard’s Nassar class patrol boats swarming the fleet like locusts.

   She wasn’t afraid of losing her Destroyer. Its R2-D2-like Phalanx and SeaRAM Close-In Weapons Systems shredded the IRGC speedboats with tungsten and vaporized them with Rolling Airframe Missiles. But even if they got danger-close, they weren’t after her ship. They were headed for the carrier.

   What she feared was far worse than death. She would have rather gone down with ship that let her fellow sailors down. Years later, she could still see the speed boats, driven by suicidal Guardsmen, explode against the carrier’s hull. Then the carrier listing, with aircraft and crew sliding off, before standing straight up like a skyscraper. And finally its plunge into the deep, sucking those Americans who’d made it off under with her.

   There was nothing she could have done, even had the destroyer – or the fleet for that matter – been under her command. But the guilt was still palpable over a decade later. She often wondered if it was why she had taken command of ship with no one aboard but her.

   That didn’t mean she wanted to lose her tanker now. And even though the sonar ping that so confused the AI did not belong to a speedboat, let alone a swarm of them, she knew immediately that it was a possibility.

   Speedboats would have actually made sense. Somali pirates still used them, even if they now were smart enough to operate them as drones. What didn’t make sense to Wexler’s sensors was the presence of a mine.

   Wexler had seen mines in her second tour in The Gulf, as a Lieutenant on an Avenger Class minesweeper operating out of Bahrain. With NATO and Emirati troops occupying the Eastern Saudi provinces, Iran was determined the West and their Sunni allies would not get a drop of oil out of the Gulf. So they mined the hell of it.

   Mines were part of the same asymmetrical Iranian war plan that produced the patrol boat swarms. Each cost no more than $10,000 but could take down billion dollar surface vessels. That never happened under Wexler’s watch.

   Lieutenant Wexler’s was in charge of her ship’s Knifefish, a specialized Bluefin-21 Unmanned Underwater Vehicle. She programmed it to hunt mines autonomously, and report back to her. It never let her down, and she never let her ship or her fleet down. The Avenger and its Knifefish were where her affinity for crewmates that lacked flesh and blood began.

   So it was, a decade later, with no small amount of guilt that Captain Wexler disengaged her tanker’s navigation system. She actually said “sorry” out loud as she shut the AI down, and began to manually steer her ship out of the path of the mine.

   It was no quick task, and not just because it was the first time Wexler had truly been at the helm of her ship. There’s a reason turning a big ship around is a metaphor for how hard it is to affect change a bloated bureaucracy, Wexler thought. Just like the international bodies that had been unable to change maritime law for the world she was living in. The same laws that prevented civilian ships like hers from being armed against pirates and other threats.

   Still, the sonar had detected the mine – and she’d recognized its distinct signature – while it was far enough away to give her all the time she needed. When her maneuver was complete, she turned to the display screen expecting to see that distance growing. Instead, she saw the mine was closing in on her.

   As if that wasn’t enough to disquiet Captain Wexler, the bridge soon became a cacophonous mixture of flashing lights and blaring klaxons as the ship’s various systems started arguing with one another. They simply could not comprehend what was chasing them. Even amidst the chaos, Wexler could understand the threat all too well.

   The underwater contact that was pursuing her vessel was no ordinary mine. It was a modified Sea Predator. Autonomous, capable of lying in wait for its prey and following it. And, once in range, more than willing to deliver a barrage of lethal self-propelled warheads.

   Wexler didn’t take any pride in her third Gulf tour. No one in the allied Navy did. By that time, The House of Saud had fallen completely, along with the other Gulf monarchies. The Gulf transformed into metaphorical gulf as well as a literal one. A string of failed states were on one side, and a weakened but still deadly Shia theocracy on the other.

   The President had determined that none of these entities would be supplied through the body of water that so many Americans had drowned in. And so it was that then-Commander Wexler – her wartime commissions on the previous tour bumping her two notches closer to her goal of Captain – went from mine-slayer to mine-layer.

   Because of their autonomy, the Sea Predators Wexler and her fellow seamen had left behind were able to hunt and kill surface vessels for years after America had withdrawn. Wexler didn’t like to think of the carnage she left in her wake. The nightmares of fishing vessels and pleasure boats being sunk by the UUVs she’d left behind were why her dream of becoming a captain could only be realized in the civilian sector.

   Years later, a multinational force led by the Indian and Chinese navies would return to The Gulf to find and disable the mines. But they never found, let alone disarmed, all of them. While minefields could be marked and sealed off, the ocean never stayed still. Currents made the Sea Predator’s kill-box a moving target that could never truly be eliminated.

   Captain Wexler imagined that it wasn’t just drift that brought the mine from the Gulf to the Horn, to her. She didn’t have to read Coleridge to know this was her albatross, and feel like she was deserving of its curse.

   But Wexler wanted to live. She didn’t want to go down with her ship. She didn’t want her ship to go down. It wasn’t out of any particular loyalty to her ever-shifting corporate paymasters. As grating as the sounds and lights flooding the bridge were, she had grown inexplicably fond of the systems behind them.

   Like the men and women she’d sailed with, the AI kept her alive and afloat. More than that, they were her raison d’etre. Even if for the majority of the voyage they were subordinate to her in name only, they were her crew. She would not let them down.

   Captain Wexler ordered her communications AI to make contact with the mine. She had programmed the mine – or one like it – and could tell it to stop pursuing her. Failing that, the AI was smart enough to brute force hack it.

   The mine did not to respond to the many frequencies Wexler and AI tried to hail it on. She remembered towards the end of the war that some officers had deliberately made the Sea Predators impossible to communicate with. In court-martials, they’d later claim it was to prevent the Russian hackers that were working with Iran from hijacking the Predators and use them in the Bering Strait.

   But that meant their orders couldn’t be rescinded. It was one of the major reasons why there was still no peace treaty between The U.S., The Islamic Republic of Iran or the various Caliphates. What kind of peace could they make with an enemy that had no ability to call off its attacks?

   Despite Wexler’s inability to disable the mine, the light and noise on the bridge began to subside. While it was slow to turn a massive ship around, once it was back on course it was able to move at a speed that a UUV simply could not match. Slowly but surely the distance between predator and prey grew. Soon the mine would lose track of her ship, and go back to its lonely patrol.

   Captain Wexler could have forwarded the ship’s log to corporate HQ in Singapore and angled for a bonus.  But something nagged at her. Why hadn’t the Sea Predator fired? Although categorized as a mine, it wasn’t the suicidal weapons platform that the Iranian speedboats were. It was capable of launching multiple self-propelled warheads.

   Perhaps it hadn’t been in range. The AI had spotted the anomaly quickly, Wexler identified the Sea Predator as a threat and course corrected almost immediately. Perhaps it had already emptied itself of its self-propelled warheads many years and nautical miles ago.

   But what if it hadn’t? What if the next captain – human or AI – didn’t share her knowledge, experience or initiative. Could she live with the consequences?

   The bridge began to erupt again, emitting more frantic warning stimuli as Captain Wexler did the unthinkable. She slowed down, allowing the predator to close in on her ship. Wexler had to literally rip wires out to disconnect systems designed to override a suicidal captain. Despite a robust entertainment suite, it wasn’t unheard of for someone to go crazy without human contact.

   Wexler was taking a risk neither her AI nor the shipping company’s shareholders would appreciate. But it was a calculated one. She didn’t intend this voyage to end with a sunken vessel. Or her death.

   The sonar was still active. She plugged in a series of personal recording devices and began downloading her ships acoustical signature over her diaries, music and podcasts. As soon as she’d copied the signature onto her devices, she hooked each one up to any speaker that wasn’t bolted down.

   Then Wexler made her way to the lifeboats. With a sole crewman, all but one were redundant. But the lifeboats were light enough that their extra weight didn’t cost the company enough fuel to risk pissing off the few skilled captains willing to take these lonesome jobs. So they remained anchored to the side, until this captain had placed a recording device and speaker in each one and set each to play on a loop.

   Captain Wexler ran back to the bridge. Her ship had slowed considerably, and if the Sea Predator wasn’t in firing range yet, it would be soon. She ordered the emergency AI to lower the lifeboats. All of them.

   The Sea Predator could have no warheads, or it could have a full payload. Saving one lifeboat for herself would mean that even if one hit, she’d have a chance at survival. But it would also mean she’d never know if she’d de-fanged the Predator she felt responsible for.

   The lifeboats soon splashed down and began to drift away from the tanker in multiple directions. Captain Wexler manually pushed her ships propulsion to the limit. Even though there was no one around, she still couldn’t bring herself to say “damn the torpedoes, full steam ahead.” But she thought it.

   It would have been appropriate. In part because the quote, attributed to Admiral Farragut in the Battle of Mobile Bay, was not actually referring to what modern naval officers would call torpedoes; in the Civil War those “torpedoes” were tethered naval mines. And in part because at that moment, The Sea Predator launched its full payload of self-propelled warheads.

   With the AI systems silenced, Captain Wexler watched in mute horror as the Predator’s warheads headed for the distinctive sound of her tankers turbine driven propellers. A sound Wexler – or any human – could not actually hear. Nevertheless, it was a sound not much different from the ones she’d programmed her Sea Predators to seek out in the Gulf.

   But then the warheads began to spread out. They had their own AI, which calculated a higher likelihood of success taking out multiple stationary targets emitting the same noise as the larger but faster moving tanker. What the warheads could not know was that those sounds were recordings, and those targets empty lifeboats.

   Captain Wexler couldn’t hear the warheads detonate behind her. She couldn’t see the explosions of wood, metal and fiberglass, or the sea spray they kicked up. All she could see were the blips on the sonar representing the warheads disappear one by one on her monitor.

   If she could have seen and heard the blasts, perhaps they would have erased the sights and sounds of the carrier she couldn’t save. She doubted it. As much as she cared for the AIs that she carefully – lovingly, even – plugged back in, neither they nor the tanker she saved could ever make up for sailors she’d seen the sea swallow that day.

   The ledger that contained the lives of the men and women her mines had taken was equally as large, if not more so. But as Wexler watched the disarmed Sea Predator slowly move towards the fringes of her sonar display, she thought that that maybe, just maybe, she’d finally put a mark down she could be proud of.

Mark Sable is a writer best known for the graphic novels Graveyard of Empires and Unthinkable, and has written Marvel and DC comics as well. He also works in film and television with experience at NBC, Fox, and Cartoon Network. He holds an MFA  from the NYU Tisch School of the Arts, a J.D. from the University of Southern California Law School, and teaches at The School of Visual Arts in New York. He can be found on twitter at @marksable.