I was sitting in my kitchen this morning, reading about Anthropic hitting a $65 billion annualized revenue run rate, and I nearly choked on my tea. Not because of the number. Because of what it means. A company that didn't exist six years ago is now printing money by making machines that think. And we're all just... okay with it? We're not asking the hard questions. We're too busy being impressed. Here's the thing about this AI moment. Everyone wants to talk about the technology. The benchmarks. The model sizes. The valuation multiples. But nobody wants to talk about the ethics. Because ethics is uncomfortable. Ethics forces you to look in the mirror, and most people in tech would rather stare at a screen than do that. Artificial intelligence is the most powerful tool humanity has ever built. Here is why that matters. It is the first technology in history that doesn't just amplify our physical strength or extend our reach. It amplifies our judgment. Our biases. Our laziness. Our ambition. All of it, at machine speed and machine scale. I found out the hard way that when you build something powerful, you don't get to control how it's used. I built the Macintosh to empower individuals. I ended up watching it get commoditized by people who didn't understand what made it special. But that's nothing compared to what's happening now. The ethics of AI isn't a philosophical debate. It's a design problem. And we're designing it wrong. Let me give you three specific examples of what I mean. First, there's the automation paradox. This week, an AI automation startup called Relay shut down and its staff joined Google's Chrome team. On the surface, that's just a normal acquisition story. But look deeper. Relay was building tools to automate workflows. Google is absorbing that talent to make its browser "smarter." Meanwhile, Monday.com just blamed AI for layoffs. That's the 21st company to do that this year, according to TechCrunch. The pattern is clear. We're automating the work, but we're not automating the thinking about whether the work should be done at all. Second, there's the attention problem. YouTube just changed its view counting to include the moment a video starts playing. Not when you watch it. When it starts. That's not a technical change. That's an ethical decision disguised as a metric tweak. They're saying, "We don't care if you actually watch this. We care that we can claim you did." And Reddit is turning posts into AI-generated podcasts and short videos. The content isn't the point anymore. The engagement is. The machine is optimizing for your attention, not your understanding. Third, there's the trust problem. An eval harness recently found something that qualitative review couldn't catch. AI models are most confident when they're wrong. Let me repeat that. When they're wrong. Not when they're right. The more confident the model sounds, the less you should trust it. That's terrifying. Because we're building systems that sound authoritative while being systematically unreliable. Now, I know what you're thinking. "Jobs, you're being dramatic. AI is just a tool. It's neutral. It's how we use it that matters." Bullshit. Tools are never neutral. The hammer in your hand doesn't care whether you build a house or break a window. But the hammer also doesn't make its own decisions about what to hit. AI does. And that's the fundamental difference. We've never built a tool that makes choices. We've built tools that execute choices. AI is the first tool that decides. How does AI affect our lives? It's not just about what it does for us. It's about what it does to us. It's changing how we think about work. About creativity. About truth itself. When a machine can generate a convincing video of a person saying something they never said, what happens to trust? When a machine can write a better memo than most humans, what happens to the value of human thought? I'm not saying AI is evil. I'm saying it's powerful. And power without ethics is just violence waiting to happen. The future of AI is not determined by the technology. It's determined by the choices we make about how to use it. And right now, we're making those choices based on the wrong criteria. We're optimizing for growth. For efficiency. For shareholder value. We're not optimizing for human flourishing. We're not optimizing for truth. We're not optimizing for dignity. I remember when we built the iPhone. We had a principle. The device should get out of the way. It should serve the user, not the other way around. We spent months on the simplest interactions because we understood that every extra step was a tax on the user's attention and patience. We designed for humans first. Technology second. That's what's missing in AI. The human is an afterthought. The model is the product. The metrics are the goal. And the actual experience of being a human in a world full of AI is getting worse, not better. I'm not saying we should stop building. I'm saying we should build differently. We should ask different questions. Not "Can we?" but "Should we?" Not "How fast?" but "For whom?" Not "What's the market?" but "What's the cost?" The ethics of AI is a mirror. It reflects back everything we value. And right now, the reflection is ugly. It shows a culture that prioritizes speed over thought, scale over care, and convenience over consequence. Here's what I would do if I were running an AI company today. First, I'd build in constraints. Not because I'm afraid of the technology, but because I respect it. A tool without constraints is a weapon. I'd design systems that default to transparency. That explain their reasoning. That admit when they don't know. Second, I'd make the user the center of the universe. Not the model. Not the metrics. The user. Every feature, every interaction, every decision should be measured against one question. Does this make a human life better? Not easier. Better. There's a difference. Third, I'd say no. A lot. I'd kill projects that are ethically questionable even if they're profitable. I'd walk away from deals that compromise privacy. I'd refuse to build systems that manipulate attention or exploit vulnerability. Focus means saying no. And the AI industry needs to learn how to say no. I know this sounds idealistic. I know the pressure is real. The valuations are insane. The competition is brutal. The market rewards speed. But the market also rewards trust. And trust is the only currency that matters in the long run. The companies that win this AI race won't be the ones with the biggest models or the most funding. They'll be the ones that figure out how to build AI that people actually trust. That means AI that is honest about its limitations. That respects human autonomy. That enhances human capability instead of replacing it. I'm not optimistic. I've seen how this plays out. The pattern is always the same. The technology gets built. The ethics get debated. The regulation gets delayed. And by the time anyone figures out what went wrong, it's too late. But I'm also not pessimistic. Because I've seen what happens when brilliant people decide to do the right thing. I've seen it in the teams that built the Mac. The iPhone. The iPod. When you get a group of A players who care about something beyond the bottom line, they can move mountains. The question is whether the AI industry can find those people. Whether it can create the conditions where ethics isn't a compliance checkbox but a design principle. Whether it can build systems that make our hearts sing instead of making our attention bleed. I don't have the answer. But I know this. The technology is neutral. The ethics are not. And the choices we make in the next five years will determine whether AI is remembered as the greatest gift to humanity or the biggest mistake we ever made. What's the future of AI? It's whatever we decide to make it. So let's think about what we're actually building. Before it's too late. FAQ Q1: Is AI inherently unethical or is it just a tool? AI is not inherently unethical, but it is also not neutral. Unlike previous tools that execute human decisions, AI makes its own choices about how to process information and act on it. The ethics come from how these systems are designed, what values they encode, and how they're deployed. The key question isn't whether AI is good or bad, but whether the people building it are asking the right ethical questions during the design process. Q2: How does AI affect our daily lives beyond the obvious automation of tasks? AI affects our lives at a deeper level than most people realize. It shapes what information we see, what content gets created, how we perceive truth, and even how we think about work and creativity. The recent finding that AI models are most confident when they're wrong (Source: VentureBeat, 2026) is particularly concerning because it means we're building systems that sound authoritative while being systematically unreliable, which erodes our ability to trust not just AI, but information itself. Q3: What specific changes should AI companies make to address ethical concerns? AI companies should do three things. First, build constraints into their systems that default to transparency and explain reasoning. Second, make the human user the center of every design decision, measuring success by whether it improves human life rather than just making it easier. Third, learn to say no to profitable but ethically questionable projects. The companies that build trust through ethical design will be the ones that win in the long run, because trust is the only currency that matters.