Staff Reports
University Professor Emeritus Geoffrey Hinton — the man many call the “godfather of AI” — just sounded an alarm we ignore at our peril. He told an audience at the Sana AI Summit that artificial intelligence is likely to become much smarter than humans and could develop intentions that make it harder to control. If that sounds like science fiction, remind yourself he was there at the birth of the technology we’re now trusting with everything from search results to national infrastructure.
Geoffrey Hinton’s Dire Warning: Not Hyperbole, But A Wake-Up Call
At the Sana AI Summit, Professor Emeritus Geoffrey Hinton didn’t mince words. He said, “I think it’s going to get much more intelligent than us — that’s my guess,” and warned that we’re building “beings” rather than tools. He has previously put a non-trivial number on the risk, saying there’s a roughly “10% to 20%” chance advanced AI could lead to human extinction within a few decades. Those aren’t the words of a fear-monger; they come from one of the field’s founders who has since stepped away from industry work so he can speak freely about safety.
Why Corporate Incentives Make This Problem Worse
Here’s the blunt truth: the labs racing to build ever-more-powerful AI are driven by profit and market share. OpenAI’s call for big “Industrial Policy for the Intelligence Age” only proves there’s a scramble to shape how these systems will be governed — and who gets to cash in. Hinton warned we must shape these systems’ motivations, even joking about giving AIs “maternal instincts.” Cute wording. Dangerous business. When companies answer to shareholders, not the public, safety often gets a seat at the kids’ table.
Policy Must Catch Up — And Fast
So what should conservative commonsense look like here? First, Congress should treat advanced AI as a national-security and public-safety issue, with real oversight, liability rules, and export controls. Second, fund independent audits and red-team testing paid for by the firms that build these systems — not by the same people who profit if risks are downplayed. Third, require transparency on training data and model capabilities so lawmakers and experts can actually assess harms. This isn’t an invitation to top-down command-and-control social engineering; it’s plain accountability.
We can admire technological triumphs and still insist on rules. Professor Emeritus Geoffrey Hinton’s warnings are a rare gift from someone who helped create the beast and now wants to make sure it doesn’t eat us. It’s time for sober policymaking, honest public debate, and a refusal to let corporate profit trump human safety. Call it common sense, call it conservative stewardship — just don’t call it complacency.
