Roundtables: Could AI really kill us all?

There's something deliciously weird about the AI extinction debate: we've created an intelligence and immediately started fearing it like an ancient god. The roundtable at MIT Technology Review (https://www.technologyreview.com/2026/09/15/1143936/roundtables-will-ai-really-kill-us-all/) dives into whether the doom-mongering from AI lab insiders represents genuine scientific concern or collective techno-anxiety. What makes this discussion so fascinating is that it's not coming from sci-fi writers or doomsday prophets — it's coming from the very people holding the soldering irons and writing the training runs. The extinction fear taps into something primal yet uniquely modern. Throughout history, we've feared gods, plagues, and nuclear annihilation — forces beyond our control. But AI is different: it's a force we created, trained, and are actively scaling up. The roundtable unpacks whether this is hubris getting its cosmic comeuppance or prudent engineering risk assessment. When employees at OpenAI, DeepMind, and Anthropic sign public statements about existential risk, they're not just being dramatic — they're engaging in the most consequential thought experiment humanity has ever run on itself. What strikes me as particularly wild is the epistemic asymmetry at play. We cannot run a control experiment with humanity — if the AI apocalypse happens, there's no "what if" scenario to compare against. This makes the debate simultaneously the most important and least empirically resolvable question in science. It's like trying to prove a negative while the positive would erase the proof. The roundtable navigates this paradox with refreshing honesty, acknowledging that the absence of evidence isn't evidence of absence when it comes to superintelligent systems. But here's the beautiful twist: even if the doomsayers are wrong, the conversation itself is valuable. By grappling with our own creation's potential to surpass us, we're forced to confront what makes us human — our values, our fragility, our desire to persist. Whether AI kills us or merely changes everything about how we live, the discussion forces us to articulate what we actually want from intelligence, consciousness, and existence. That's the real weirdness: we're using the threat of our own obsolescence to finally figure out who we are.