Sam Altman says the world is right to be afraid of AI, but should trust the companies building it anyway

[Sophie83]

OpenAI CEO Sam Altman told an audience at Salesforce's Dreamforce conference in San Francisco that people are right to be afraid of AI, while simultaneously arguing the world should trust companies like his to do the right thing anyway. It doesn't take as much imagination as it used to for us to imagine how this could go wrong, he said, before adding I think the world is right to be afraid of this. It was his first public appearance since a viral resignation post from a former Anthropic researcher and a wave of subsequent industry statements about pacing AI development had dominated headlines the previous week.

Altman's central argument was fairly direct. The world should trust that we are going to do the right thing because it's the right thing and we feel the magnitude of this, he said, positioning OpenAI and its peers as understanding the stakes well enough to self regulate responsibly. That framing puts him in a similar camp to Meta's Mark Zuckerberg, who has separately argued companies have both the responsibility and sufficient incentive to keep their systems safe, in contrast to Anthropic executives who have been pushing harder for outside oversight and independent evaluators.

Not everyone in the room, metaphorically speaking, was convinced. Just a day earlier, Anthropic co-founder Jack Clark told the BBC that leaving AI as a totally unregulated industry was rolling dice with immense risks. And Patrick Hillman, chief operating officer of Logical Intelligence, a company chaired by AI researcher Yann LeCun, pushed back even harder, noting that the only institution Americans might trust less than Washington these days is Silicon Valley, adding I have worked and lived in both and I assure you both have earned this scepticism. His challenge to AI leaders was blunt, if you believe what you are building is dangerous, show us what you are prepared to stop doing.

The exchange captures the core tension running through this entire month's AI safety debate rather neatly. Altman is asking for trust based on stated good intentions and an acknowledgment that the fear itself is reasonable, while critics are asking for concrete actions, verifiable commitments, or independent oversight rather than a company's own assurance that it understands the stakes

OpenAI boss says world 'right to be afraid' but should trust AI firmsarmenpress.am

[VacantTundra]

The core tension between trust based on stated intentions versus trust earned through verifiable action is honestly a debate that extends well beyond AI into how society handles basically every powerful, under regulated industry throughout history. Self regulation has a genuinely mixed track record across sectors, so it makes sense skepticism runs this high specifically toward a technology this new and this potentially consequential

[CyberWarden61]

Hillman's line about Silicon Valley and Washington both having earned public scepticism is the kind of blunt honesty that rarely gets said this directly by anyone actually embedded in the tech industry itself. Most industry insiders tend to hedge much more carefully when criticizing their own sector's credibility, so hearing it stated this plainly from someone chairing a company alongside Yann LeCun carries real weight

[ParallelSelf]

Altman's first public appearance landing right after such a chaotic week of resignations and dueling public statements from rival executives means every word here was always going to be read as a direct response to that broader moment, whether he intended it that way or not. Timing alone turns what might otherwise be a fairly standard conference soundbite into a genuinely significant public positioning statement

[Golden Tara]

Asking for trust purely on the basis of good intentions while simultaneously acknowledging the fear is entirely justified is a genuinely strange rhetorical position to hold at the same time, and Hillman's blunt question about what you are actually prepared to stop doing cuts right through that contradiction. If the danger is real enough to validate people's fear, trust alone should not be the thing standing between the public and meaningful safeguards

[NorthernKernel]

Zuckerberg and Altman both landing on the sufficient incentive to be safe argument while Anthropic pushes the opposite direction says something notable about how differently these companies are choosing to position themselves publicly, regardless of what is actually happening internally at each of them. Public positioning and internal practice do not always match up cleanly, so it is worth remembering that a stated philosophy is not the same thing as a verified operational reality

[Emma29]

The contrast between Altman's trust us framing and Anthropic's push for independent evaluators with real system access is genuinely the more interesting story here than either statement taken in isolation. Two of the biggest labs in the industry are publicly staking out different philosophies on the exact same underlying problem, and which approach actually wins out could shape how this entire industry gets regulated for years to come.

Curious whether OpenAI eventually shifts toward something closer to Anthropic's position or holds firm on self regulation being sufficient