Google’s AI Shift: The End of ‘Don’t Be Evil’?
05:40 Share

Welcome back to the AI podcast. I'm Eli, your host, and today AI is moving faster than ever and the latest developments might just change everything. So let's break it down. Imagine you're on a battlefield, right?

The tension is crazy high every second counts and decisions have to be made like that. Now imagine AI stepping in to like speed up that whole process of, you know, IDing threats, tracking, and then taking them out. That's basically what's happening, folks. The Pentagon's AI chief straight up admitted some companies are providing these AI models to accelerate the military's kill chain, as they call it.

So that's what we're diving into today, this whole relationship between AI developers and the military and all the ethical stuff that comes with bringing AI onto the battlefield. It's pretty wild when you think about how much things have changed. It wasn't that long ago Google was super outspoken about AI and warfare. They even had this whole pledge on their website saying they've never used their tech for weapons or surveillance. But get this. Recently, they quietly scrubbed that promise from their website. Just proof. God. They didn't.

They did put out a new blog post about responsible AI, though, which makes you think, like, what does responsible AI even mean? I mean, we're talking about the military, right? Like, where do you even draw the line? The thing is, Google's still saying they're committed to human rights and all that and avoiding harmful outcomes. But how do you do that when AI is being used in situations?

where it could be like making life or death decisions right there on the battlefield. It's like walking a tightrope. And Google's not the only one trying to figure this out. Companies like OpenAI and Anthropic are also jumping into this military AI thing. They're selling software to the Pentagon, but they're super careful to say that their AI is just for making things more efficient, not for actually pulling the trigger. But here's where it gets tricky. The Pentagon themselves are saying that AI is giving them a huge advantage when it comes to that whole kill chain process.

So how do you separate helping with strategy from enabling actual lethal action? Tough question. Are we talking about AI as like a strategic advisor just helping humans make better decisions? Or is it becoming something more, something that plays a more active role in war? Because some experts are even saying that the U.S. military already has autonomous weapons, systems that are making life and death decisions without a human giving the order.

Or as the Pentagon says, they'd never let a weapon system run totally on its own. They say there'll always be a human in a loop, especially when it comes to using force.

But it makes you wonder, what does human in the loop even really mean? When you have AI crunching data super fast and suggesting what to do, is there even enough time for a human to really think things through? In those super intense moments, you know when it really matters. This whole thing gets even more complicated when you start talking about like AI coding agents. You know those AI systems that can write and debug code themselves? Imagine that being used to develop military software to kind of make you wonder who's really responsible.

when those programs start doing things out in the real world. And it's not just coding agents either. We're seeing more and more automation everywhere, right? Self-driving cars, even self-firing turrets, believe it or not. So where's the line between automation and true autonomy? If we can't even agree on what autonomy means, how can we set up any ethical rules for military AI?

This isn't just some theoretical thing either. Remember last year when those employees at Amazon and Google were protesting about their company's military contracts? Well, there's talk of something similar happening in the AI community now. Some researchers are saying they want to work directly with the military. They say it's the only way to make sure AI is being developed responsibly for defense. But others are drawing a hard line. They're like, no, my AI will never be used to hurt anyone. But even those lines can get blurry. Like take Anthropic, for example.

They're a big AI company and they have this strict policy against using their models for anything that could harm people or cause loss of life. Sounds pretty clear, right? Well, here's the thing. Anthropic is also working with Palantir, that company that provides data analysis tools to the military. And it doesn't stop there. They're also working with defense contractors like Lockheed Martin and Booz Allen. So it makes you think, are they really sticking to their principles? Or is this case of do as I say, not as I do?

And then there's OpenAI, another big name in AI. They've got this deal with Andoril. They're a defense tech company known for their work on autonomous drones. It feels like a lot of these AI companies are trying to have it both ways. They want to work with the military, but they also want to look like they're on the right side of history. I get both sides of the argument, though. AI could totally change the game in defense. It might even save lives in some situations. But it also opens up this whole Pandora's box of ethical problems that we can't just ignore.

It all comes back to this one question. Who's really in control when AI is involved in military operations? Are we okay with machines having that much power, even if it's of national security, because things could go wrong? And we're talking about weapons systems. Going wrong could mean something really, really bad. Now, I'm not saying we need to ditch AI and defense entirely. There's got to be a way to use it for good and keep humans in control.

Maybe it's all about being transparent, having these open conversations and setting clear rules. Maybe we need to get more people involved in the conversation too, not just the tech companies and the military, but like ethicists, philosophers, even regular people. It's a super complicated issue, but he's something we have to figure out because we're talking about the future of war here, maybe even the future of humanity.

How do we make sure humans are still making important calls? The human judgment and compassion are still at the heart of it all. That's the big question we need to answer. As AI becomes a bigger part of military stuff, it's something to think about for sure. And that's all the time we have for today. Keep those lines curious.