I am part of multiple startup founder communities in the UK. Job boards, networking meetups, communities, practically all the ways that founders try to help each other out. During the initial days of Occam, when I was recruiting, I thought I’d send out feelers that we were looking to hire. I posted on one of these job board groups that I was looking for software engineers to join my team. I said explicitly that this is a defence job, and that we’d be deploying technology used on the front line, and that lives would directly be impacted by our engineering decisions. I got something like 14 puke emoji reactions to that post. A few people even said they hoped we wouldn’t find anyone.

That stuck with me. And it wasn’t a one-off. Across the UK and Europe, when I tell people I am working on AI for defence, or building AI pilots for autonomous drones, I get reactions ranging from subtle derision and admonition to outright judgement about what that kind of work represents, as if I’m choosing to indulge in mechanised violence at scale.

It feels a bit off

I never really understood this judgement as a person from India. I would go so far as to say that we in India learn certain lessons growing up that feel more immediate than they do here. We learn very young that the locks on your doors aren’t just for show, there are robbers in the neighbourhood. As India’s late former President Dr APJ Abdul Kalam very succinctly said in an interview, each nation needs peace so that it can engage in nation-building activities. That peace only comes from strength. Strength respects strength and nothing else. Indians have the largest all-volunteer army in the world. Even today in India, we say Jai Jawan and Jai Kisaan, which translates to Glory to the farmers who give us food and the soldiers who protect us. From what I’ve seen, the idea that “defence is bad” doesn’t really come up in the same way in India.

Yet, this anti-defence sentiment comes up often enough in Europe that it’s hard to ignore. Which is why I decided to write this post. I’ve been in Europe for almost 8 years now, and therefore feel a semblance of attachment to this continent. I may be no native, but I still feel compelled to write this down from the perspective I’ve had growing up elsewhere.

I think people forget their history.

Europe learned this lesson multiple times in the past. In my opinion, the younger generation of today, who are further removed from the wars of the past, are the ones who don’t quite see the role of defence as clearly. The counterfactual is also possible - that people are so disgusted by the idea of war precisely because they’ve seen too much of it. That part is understandable.

If you are having a lot of fights with your partner, you become more conscious of it and try harder to avoid it. You teach your children not to raise their tone, not to get into fights. You also teach them that when mummy and daddy fight, they are wrong - just so the kids don’t grow up normalising it. Similarly, I think the general sentiment of decrying the very ethical basis for war, defence, or anything related to the army has taken hold of public discourse in Europe for a long time. And Europe was able to operate this way for a long time. It had America to cater to its defence needs, the Middle East and Russia to cater to its oil needs, and China to cater to its manufacturing needs. Which led Europe to focus more on being a place that prioritised quality of life and the finer aspects of society. Fair enough.

However, Karl Popper has this idea called the paradox of tolerance. If a society extends tolerance to those who are fundamentally intolerant, the intolerant will eventually undermine both the tolerant and tolerance itself. At a high level, this starts to look less like a permanent state, and more like a temporary equilibrium. Over time, a belief starts to take hold - that maybe the world is a nice place. People can be trusted. The world is not out there to rip you off. We can drop our guns, trust everyone, be nice and share with everyone. As much as I like believing that, I still lock my doors at night.

Strength respects strength

There’s a parable about the Buddha that has always stuck with me. More folklore than canon so wouldn’t be surpised if this doesn’t have an official source. Buddha once went to Vaishali and spoke to its people about peace and non-violence. He taught that clinging to weapons and endless conflict only brings more suffering. The people of Vaishali, known for their free-spirited, republican ways and love of debate, took his words deeply to heart. They decided to lay down their arms and embrace a life of peace.

Then he travelled to Magadha and delivered the same message. But Magadha was different. It was a powerful, centralized kingdom with a more militaristic spirit. They listened politely enough, but disagreed and ultimately carried arming up.

Later, when Magadha realised that Vaishali had disarmed while they remained strong, they saw an opportunity and conquered the city without much resistance. It’s said that when the news reached the Buddha, he frowned. Probably one of the few times in his life. The story is often interpreted as a moment where his teachings, applied asymmetrically, left the peaceful vulnerable. In a world where not everyone plays by the same rules, peace ultimately rests on strength.

Deterrence is a necessity

That doesn’t mean people need to be war mongers. I’m not trying to glorify violence. The way I see it, it’s as important as any other profession and institution of society. You can have the bankers and the farmers, and the mechanics and the carpenters. And you can have the defenders. It’s not wrong. It’s a necessary element. Not even a necessary evil. Just necessary. A software engineer working on defence is helping secure a system that everything else depends on. A forward deployed engineer talking to soldiers is solving real problems in real environments. It’s not fundamentally different from any other critical system we rely on. I don’t see working in defence as inherently unethical. It doesn’t have to be seen as hypocritical. And it doesn’t require dismissing people who choose to work in it.

There’s a caveat here though. I am talking about defending whats yours. Not justifying attacking others. In practice, the line isn’t always clean. I’m not naive enough to claim that. Nor am I claiming that people won’t build an industrial complex and use the machinations of capitalism to profit endlessly from the fear of rain by selling umbrellas. That’s a different article. What I am saying is, the ability to defend usually implies the ability to impose costs on an attacker. But intent still matters. There’s a difference between building capability to protect yourself and using it to exploit others. At a minimum, I don’t think working in defence automatically puts someone on the wrong side of that line.

Lethally autonomous weapon systems

Coming to the use of LAWS (Lethally Autonomous Weapon Systems), often called “killer robots”. I’d be remiss if I didnt talk about it in this article.

I’ve heard the argument for “defence is fine, just don’t make killer robots”. That’s a position I can at least understand. People feel the line should be drawn at giving machines agency. I’ve been asked if by giving AI agency, I am building the Terminator, or VIKI from I, Robot.

The case against LAWS is that we shouldn’t let machines have more than 2 out of 3 - detect, decide, destroy. There’s also the idea of humans being “in the loop” vs “on the loop”. The former means humans are part of the decision-making process, whereas in the latter the decision is made by the AI with a human supervising. I’ve lived this problem throughout my PhD from the work by Sheridan and Verplank to my own work on autonomy switching and mixed-initiative systems.

And I don’t think it is actually possible to get a human out of the loop in the first place. Where does decision-making actually sit? If I set a thermostat to maintain 21°C, the system decides when to turn heating on or off. But the objective function, constraints, and acceptable actions were all defined by me. The “decision” is delegated, not originated.

As people, we are used to being responsible for both the decisions we make and their consequences. When you are having an argument with your partner, you are both making decisions and carrying responsibility for them. A member of parliament legislating laws is executing decisions towards what they believe is a morally good outcome. So we instinctively bind decision-making and moral accountability together.

What technology allows us to do is decouple them. We can delegate decision-making to a system within a defined framework, while still being fully accountable for the outcomes. A soldier pulling a trigger carries intent. The weapon does not. The same applies whether it’s an AI drone, an autonomous turret, a guided missile, or any other system. The question is who is responsible for sanctioning the violence not whether it is mechanised or manual. Mistakes by machines trace back to how humans designed and deployed them. Just like your boss is responsible for a bad decision they delegated to you, responsibility doesn’t disappear when decision-making is delegated.

Parting Thoughts

This question of agency and responsibility isn’t new. The Bhagavad Gita is essentially about this action under moral ambiguity. Acting when the outcome is uncomfortable, but the responsibility is still yours. (Although this doesn’t mean there’s no responsibility on the designer of the defence machinery. That however is another article.)

Anecdotally, I’ve seen this in a more grounded way as well. My mum has a very blunt way of putting it that ethics, in practice, often looks like the decisions a knife in the hands of a butcher has to make. The knife doesn’t decide. The hand does. That framing has always stuck with me. Because it brings the question back to where it actually belongs not on the tool, but on the person using it. The question of whether you have more agency or less agency than you think, a valid disagreement. Happy to discuss that over a coffee.