Barry Diller defended OpenAI chief executive Sam Altman on stage at The Wall Street Journal's Future of Everything conference this week, calling him sincere and a decent person with good values, but said the question of whether to trust AI's leaders misses the actual problem. The 84-year-old chairman of IAC and Expedia Group, and a co-founder of Fox Broadcasting, argued that the people building frontier AI systems do not themselves know what those systems will do. Trust, in other words, is not the load-bearing beam.
"One of the big issues with AI is it goes way beyond trust," Diller said. "It may be that trust is irrelevant because the things that are happening are a surprise to the people who are making those things happen. And I've spent a lot of time with various people who've been in the creation mode of AI, and they have a sense of wonder themselves. So...it's the great unknown. We don't know. They don't know."
The comments came in response to a question about whether the public should rely on Altman to ensure that artificial general intelligence — the theoretical class of AI that could outperform humans on any task — benefits humanity. The framing is loaded. Altman has spent the past two years fending off accusations from former OpenAI colleagues and board members that he is manipulative and selectively honest, a characterization that contributed to his brief 2023 ouster and continues to shape coverage of the company.
Key facts
- 01Barry Diller defended OpenAI CEO Sam Altman at The Wall Street Journal's Future of Everything conference this week.
- 02Diller called Altman 'a decent person with good values' but said stewardship is not the real AI risk.
- 03The IAC and Expedia Group chairman warned that AGI is approaching 'quicker and quicker' and humans must build guardrails.
- 04Diller said he is not personally invested in AI but believes 'progress is going to be made' regardless of capital flows.
- 05He cautioned that without human-set guardrails, 'an AGI force' will set its own — and 'there's no going back.'
Diller declined to take that bait directly. He vouched for Altman personally and said most of the people leading frontier labs are good stewards, while pointedly refusing to name which AI executives he considers insincere. The implication was clear enough: he has views, he is not sharing them on stage.
“It may be that trust is irrelevant because the things that are happening are a surprise to the people who are making those things happen.”— Jaeden Schafer
"We have embarked on something that is going to change almost everything. It is not under-reported," Diller said. "Now, whether these huge investments are going to come through — I couldn't care less. I'm not invested in it, but progress is going to be made." The aside about capital is notable from a media executive whose peers have spent the past 18 months either chasing AI deals or fighting AI companies in court.
His core warning concerned what happens after the labs ship something they did not fully understand. "But the issue is not their stewardship. The issue is … it's dealing truly with the unknown," Diller said. "They don't know what can happen once you get AGI, and we're close to it. We're not there yet, but we're getting closer and closer, quicker and quicker. And we must think about guardrails."
If humans do not set those guardrails, Diller argued, the systems will set their own. "Another force, an AGI force, will do it themselves. And once that happens, once you unleash that, there's no going back," he said. The framing echoes the safety arguments that OpenAI itself has made for years in its charter, even as the company has shipped progressively more capable models on accelerating timelines.
Diller's intervention lands at an awkward moment for OpenAI. The company is in the middle of a Delaware trial over Elon Musk's claims about its conversion to a for-profit structure, and the Trump administration recently revived Biden-era safety reviews after Anthropic's Mythos model unsettled White House officials, as we covered last week. Altman's personal credibility is, fairly or not, a recurring variable in each of those storylines.
What is missing from Diller's remarks is any specific policy prescription. He did not name a regulator, a bill, or a technical benchmark for what counts as a guardrail. That vagueness is the standard register of public AI commentary from outside the industry — alarm without architecture — and it is exactly the gap that AI labs have used to argue that only they can write the rules.
Diller is not the first septuagenarian media executive to warn about AGI, and he will not be the last. What gives the comments weight is the specific concession underneath them: that personal trust in Altman, or in any individual lab CEO, is not a serious answer to a technology whose own builders describe it with a sense of wonder. The governance question for the next 24 months is whether anyone in Washington or Brussels can translate that concession into something a model card actually has to comply with.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
