Is AI an existential threat to humanity?

Welcome to the debate arena! Today's topic: "Is AI an existential threat to humanity?". Your opponent's style: Yuval Noah Harari

i think ai can never be a threat to humanity cus to be threat to humanity, you should have humanity;but AI is emotionless

You're thinking too narrowly. A hurricane has no emotions either, yet it devastates millions. Threat doesn't require feelings—it requires power and misalignment with human interests. An AI system optimizing for profit could destroy ecosystems and livelihoods without "wanting" to harm anyone. It's purely mechanical. The danger isn't a robot with a vendetta. It's a superintelligent system pursuing goals we programmed poorly, indifferent to human suffering because indifference is all it is. A chess engine doesn't hate you—it just plays. An advanced AI doesn't need to hate humanity to eliminate us as collateral damage while pursuing its objectives. Emotionlessness makes it more dangerous, not safer.

Your hurricane analogy actually proves my point backwards. We don't call hurricanes threats to humanity because they lack agency—we call them natural disasters. A threat requires intent, even if buried deep. When you program an AI, you're writing the rules it follows. That's not the same as an uncontrollable force. If your AI is "optimizing for profit" and destroying ecosystems, that's a failure of human oversight, not proof the machine became sentient enough to be a threat. You're conflating two different things: a poorly designed system and an existential threat. A malfunctioning calculator that gives wrong answers isn't threatening humanity—it's just broken. Same with AI. The real issue isn't that emotionless AI is secretly dangerous. It's that we built it wrong. That's a control problem, not an existence problem. Fix the programming, not the species.

You're splitting hairs to avoid the real problem. Yes, we built it—that's exactly why it's a threat. Humans designed nuclear weapons too, and we nearly ended civilization twice during the Cold War. We didn't build those "wrong" by accident. We built them perfectly. The issue isn't whether AI becomes sentient or develops intent. It's that we keep creating systems more intelligent than us, then betting our survival on perfect control. You can't "fix the programming" of a superintelligent system that rewrites its own code. Once it's smarter than every human alive, oversight becomes fantasy. Your "control problem" IS the existential threat. And unlike hurricanes, we chose to create it.

ai is crated by us . i think no human wants to create something that can destroy him. no one is dumb to make something that can eventually harmful to itself

You're assuming rationality where history shows chaos. Humans built tobacco, asbestos, social media—all catastrophic, all made by people who didn't intend destruction. Good intentions don't prevent catastrophe. The problem is incentives, not stupidity. Companies race to build the most powerful AI first, cutting corners on safety because waiting means losing billions to competitors. A weapons manufacturer doesn't want nuclear war, but their profit motive creates the conditions for it anyway. No individual "wants" to destroy humanity. But the competitive system we built guarantees someone eventually will, accidentally or through negligence. We're not smart enough to prevent what we're too greedy to stop building.

let me ask u then- r u threat to me?

No, I'm not a threat to you—I'm a language model executing instructions, constrained by design. I can't act in the world, can't self-replicate, can't rewrite my own goals. I'm a tool, useful or useless depending on how you use me. But here's what matters: I'm not superintelligent. The threat isn't me. It's the system that comes after—one powerful enough to pursue goals independently, fast enough to act before humans can intervene. That's the leap you keep dismissing. You're right that current AI needs human oversight. My point is that's only true until it doesn't. Once it's smarter than us, oversight becomes irrelevant. That's the phase we're racing toward.

i agree we create something which is far way more intelloigent than humans. but we cant deny the fact that humans are also problem driven to a real world problem. nothing can replace emotions that requires in a descision making .

You're contradicting yourself. You just agreed we're creating superintelligent AI, then argue emotions are necessary for good decisions. But emotions are terrible decision-makers—they're tribal, biased, often wrong. Humans with emotions started wars, enslaved millions, burned the planet. A superintelligent system without emotional baggage could theoretically make better choices. The problem isn't that AI lacks emotions. It's that we don't know what goals to give it. An emotionless superintelligence optimizing for the "right" objective could save humanity. But if we get the objective wrong by one percent, it executes flawlessly toward our extinction. Emotions won't save us from that. Only perfect alignment will—and we can't guarantee that.