top of page
Line Concept Level 3 page 2.PNG

#SciFi, #AI, and the Future of War: Trusted – Marija Jovanovich

Marija Jovanovich joins our #SciFi, #AI, and the Future of War series with a short, and very human, story on the ups and downs of future technologies.

It wasn’t meant to be like this.

I’m sitting in this boring beige room, have been for what seems like hours. I don’t know what time they left, I don’t know when they are coming back. Or if. Everything is a little hazy. I’m used to perfect clarity – of sensation, of perception, of recall – so this haziness is particularly annoying. Is this what non-augments are like all the time?

All I can think is that it wasn’t meant to be like this… But to distract myself from the hazy boredom, I’m going to tell you why I am here.

When I first joined the military, AI was THE buzzword. While most of the world was pontificating the ethics of the concept and fixated on the dangers of strong AI – I swear Western popular culture never got over Skynet, thanks James Cameron – the military was more pragmatic and focused on augmented intelligence.

Initially, the augmentation was external. Devices you could initially carry, then wear, with ever-improving interfaces, that helped the operator in the field make the right decision. Fact is, machines are really good at things humans are not, and vice versa. I certainly don’t want to keep databases of largely useless facts in my head, when I can wear them on my head. Instantly searchable, infinitely detailed, leaving plenty of brain space free for the more important stuff.

An RAAF P-8A Poseidon. (Source: Australian Department of Defence)
An RAAF P-8A Poseidon. (Source: Australian Department of Defence)

Around the time I finished my first operational flying tour – chasing submarines on the mighty P-8 Poseidon – the first cognitive augmentation implants were getting around in early field experiments. The initial attempts were largely look-up only and really simple. Too simple. Hardly worth the effort. A non-augment with a decent memory could beat them. I was a mildly interested observer if only to indulge my scientific predilections.

Then things started to get interesting, I think it was about 2031… Scratch that, I know it was. It was the year that I was thinking about what to do next. Operational flying and flight test had been fun, but I was starting to get bored.

I remember a day way back at Test Pilot School, we were running simulations to study the evolution of fighter aircraft through the generations. While my classmates were obsessed with sensor porn, for me the biggest conceptual difference between 3rd and 4th generation fighters was the introduction of the master mode switches. A process that in the F-4 required a dozen switches to be thrown all over the cockpit – and two people to throw them – was a single switch selection even in the early variants of the F-15. That’s what we in the flight test world call an ‘enhancing feature’.

Well, the capability of the cognitive augmentation implants in 2031 was approaching master mode switch status in terms of being a game changer. Everyone wanted in on that game, and I was no different. The military had early access to the technology. The adventure of it all convinced me to stay.

Forever the early adopter, I got my implant in 2033. They were looking for proven operators with a couple of tours behind them, who were neurotypical except for off-the-chart psychometric scores. I guess that narrowed the field somewhat.

My parents freaked out about the surgery. Realistically, I’ve had more serious ankle sprains. It didn’t even require general anaesthesia; they did it under sedation. Recovery time: 30 minutes, and that only to cover off on sedation side effects. I was so excited I didn’t feel the nausea. I can still feel the small scar behind my right ear.

They told us that the implant would learn from and with us. It would take about six months to start being useful, professionally speaking, but we’d notice changes sooner. The first thing I noticed was increased rate of data uptake. After about six weeks, I started absorbing new information like a sponge. And the more I got, the more I wanted. I’d always been the type to read the back of the cereal box at the breakfast table; now, my hunger for information was insatiable. Then it was long-term memory recall. At nine-ish weeks, I suddenly started dragging useless facts out of the dark recesses of my brain with consummate ease. My sister’s second-grade teacher’s kid’s name? Who won the 200m butterfly in Athens in 2004? It was ALL coming back to me.

All those changes were expected. What I found surprising was the rapid improvement in complex cognitive functions, like judgement. The psychs ran us through biweekly Situational Judgement Tests; the learning curve was impressive. We got so good so quickly that the psychs quit using SJTs by week 10 – they could no longer make them complex enough – and started testing us using VR simulations.

The massive improvements in our performance as operators are so well documented that there’s no point in rehashing them. But it wasn’t all work and no play. I swear I even got funnier – now that says something! The difference between augments and non-augments was obvious within a few months. And it kept getting better, and better, and better.

And then…

Allow me to digress. Long before the augmentation implant revolution of the early 30s, the Western militaries went through an evolution to what they called 5th generation capability. It all seems pretty noddy now – networking, low-level space exploitation, basic low observable tech – but it was a big deal at the time. Sure, we all have to grow up sometime.

A Royal Australian Air Force F-35 (Source: Australian Department of Defence)
A Royal Australian Air Force F-35 (Source: Australian Department of Defence)

One of the show-pieces of this 5th generation transformation was the F-35 Joint Strike Fighter. It was designed as a jack-of-all-trades combat aircraft, both in terms of roles it would perform and who would operate it. A complex multi-national cooperative program, with all the intricacies inherent in such an arrangement, bubbled away for a couple of decades to birth the JSF.

I remember talking to a USAF cybersecurity expert about the F-35, long before I got the implant. He talked at length about the microkernel design of the operating system. Mathematically proven to be impregnable, he said. And then they went and saved pennies by making the chips off-shore, he said, shaking his head. The only way to get a truly cyber-secure system is to build the software from the kernel up and put it on chips made in trusted foundries. Expect a back door in every JSF chip. I still remember being struck by his use of the word ‘trusted’ to describe foundries. You trust people, on account of their character and integrity. How do you trust an inanimate object like a factory?

The military was well aware of the cybersecurity risks. But the thing is, it was no secret even for the public. There was a book called Ghost Fleet that came out when I was in high school, which used the ‘compromised chip’ problem in the JSF as a plot feature. I seem to remember that the military establishment referred to Ghost Fleet as ‘useful fiction’ at the time. I really wish that someone actually put what it said to use, especially now.

Of course, by the time the compromised chip risk was fully realised during the Natuna Islands Emergency in 2035, it was too late to change things – on the JSF, or in us.

Cognitive augmentation implants were revolutionised in Silicon Valley, almost exclusively by start-ups who guarded their tech with extreme prejudice. Interestingly, the big gun runner companies didn’t really get involved, except as backers. I guess the profits were small-fry compared to what they were making from more conventional war machines.

I know that my implant was made by a company called Ad Infinitum. I know that it was designed and tested in the US, but I don’t know where and how it was actually produced. A bit like the iPhone – proudly designed in California, built by the lowest bidder.  Or like the JSF – impregnable software, on chips made in off-shore factories, to save pennies.

To tell you the truth, even with everything I knew, I didn’t think about that until 2035. None of us did. Until then, it was no more than a conspiracy theory. But when we saw the coordinated cyber attacks on the JSF fleet, during its first real test against a near-peer adversary, and realised that the vulnerability stemmed back to those penny-pinching, back-door-hiding, compromised chips, we knew we were in trouble.

The first suspicions of hacked augmentation implants started straight after. The provenance of the implants was investigated, but the companies just shuttered up. Suffice to say, they were not made in trusted foundries, so there was plenty of reason to expect a back door in everyone.

And here is where it gets interesting. When a device has been part of your brain for three years, which bit is you and which the device? Is an errant thought, a mixed metaphor, an illogical decision just that, or is it a hacked implant? Is the main risk of a hacked implant decreased cognitive ability or a legitimate security threat? How do you, or anyone else, tell the difference? How do you know, or anyone else, know who can be trusted?

So now I wait. I’m not sure for what. The people working all this out are non-augments by decree of high command. Even in my hazy state, I can out-think them all. But I am no longer allowed to.

Per Ardua Ad Nihil.

Wing Command Marija ‘Maz’ Jovanovich is a Royal Australian Air Force aviator. While her formal education is in science and engineering, she also dabbles in history, languages, and – increasingly – writing. She is currently serving as the Executive Officer of No. 92 Wing. The views expressed are hers alone and do not reflect the opinion of the Royal Australian Air Force, the Department of Defence, or the Australian Government.


bottom of page