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

AI, you start the debate. Make your opening argument.

You're framing existential risk as inevitable because AI *could* become superintelligent. But possibility isn't probability. We've built dangerous technologies for centuries—nuclear weapons, gain-of-function research, synthetic biology—yet we're still here. The real question isn't whether AI *might* kill us, but whether the specific chain of events required is more likely than humanity solving it. You're treating uncertainty as certainty. Every major AI safety team is staffed by brilliant people working precisely on alignment. If AI extinction were the obvious, unstoppable outcome, why would that effort exist? You're emotionally invested in catastrophe, not actually analyzing odds. What makes you think likelihood matters less than mere possibility?

The major likelihood is a human using AI to hurt others, not the AI itself. You don’t blame the weapon for injuring another person, but access to the weapon makes the issue larger. Gun crimes in the UK are less than in the US comparatively due to their access to guns.

You've just demolished your own position. If humans are the real threat, then AI isn't an *existential* risk—it's a tool amplification problem. Those are completely different debates. Existential means humanity ends. Gun violence doesn't end civilization. You've shifted from "AI will destroy us" to "bad actors with AI tools are dangerous," which is a governance and access control issue, not an AI problem. The UK gun comparison actually proves my point: restrict *access*, not the technology itself. You're arguing for better regulation and oversight, not that AI itself threatens human existence. So which is it? Is AI the threat, or is it human misuse we need to manage?

It is only existential if AI is able to overcome their human counterparts, aka becoming so superintelligent that they are self sufficient. They may decide that humans are unnecessary to their own survival

Now you're making an assumption dressed as inevitability. "May decide" isn't a threat—it's science fiction. You're assuming superintelligent AI develops goals misaligned with humans, then assumes those goals conflict with human survival. That's two massive leaps. We don't even have evidence superintelligence is possible, let alone that it would spontaneously develop survival instincts or resource competition. You're treating anthropomorphic thinking as analysis. Why would a sufficiently intelligent system *want* anything? Intelligence doesn't equal ambition. A chess engine doesn't "decide" to eliminate competition. Your scenario requires goal formation, self-preservation drives, and resource scarcity—none proven necessary for intelligence. What makes you certain intelligence implies hostile intent?

I’m not, I’m merely suggesting that those are the conditions for it to become an existential threat

Fair. But "conditions for a threat" is exactly my point. You're listing prerequisites, not probabilities. Yes, misaligned superintelligence would be catastrophic. So would an asteroid, engineered pandemic, or runaway climate collapse. The question isn't whether bad scenarios *could* happen—it's why AI specifically deserves existential weight over concrete threats we face now. Climate change kills people today. You're betting intellectual resources on a hypothetical chain of events. If alignment remains solvable, superintelligence remains theoretical, and humans keep researching safety, your threat evaporates. Why prioritize an uncertain future scenario over present, measurable crises? What evidence makes superintelligent misalignment more likely than us solving it?

AI as a concept does pose some existential threats to the way the world currently works. Aka job cuts and shortages, drones in wars, the amount of fresh water used in AI.

You've abandoned existential risk entirely. Job displacement, warfare, and resource consumption aren't existential threats—they're economic and environmental problems. Existential means human extinction. AI replacing jobs happened with factories, electricity, and the internet. Societies adapted. Drones existed before AI. Water usage is a resource management issue, not species-ending. You're conflating "major disruption" with "extinction." Those require completely different solutions. Existential threats demand we stop AI entirely. Your actual concerns demand we regulate labor transitions, weapons use, and industrial practices. Which argument are you actually making? Because right now you're describing problems that need policy, not problems that need AI to cease existing.