Billionaire media executive Barry Diller knows Sam Altman well enough to vouch for him. He calls the OpenAI chief sincere. A decent person with good values. Yet onstage at The Wall Street Journal’s Future of Everything conference this week, Diller delivered a blunt message. Trust is irrelevant. Not because Altman lacks it. Because artificial general intelligence operates in territory no one controls.
The comments come at a charged moment. Weeks earlier, The New Yorker published a detailed examination of Altman’s leadership. Former colleagues and board members described patterns of misrepresentation. Memos from Ilya Sutskever alleged a consistent pattern of lying to executives and the board about safety protocols. Helen Toner and Tasha McCauley, who left the board after Altman’s brief ouster in 2023, questioned whether he could be trusted with technology that might reshape humanity. Dario Amodei, who departed to found Anthropic, concluded the problem with OpenAI was Altman himself.
Diller brushed those concerns aside. He has spent time with Altman and other AI builders. He believes most leaders racing forward act as responsible stewards. But stewardship alone cannot address what lies ahead. “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.”
Short sentence. Direct. The creators themselves express wonder at their own outputs. They admit uncertainty. That admission shifts the debate. Personal character matters less when the systems produce results their makers did not anticipate.
Diller, chairman of IAC and Expedia Group and co-founder of Fox Broadcasting, does not invest in these companies. He claims no financial stake. His perspective carries weight precisely because of that distance. “We have embarked on something that is going to change almost everything,” he added. “It is not under-reported. 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.”
And progress accelerates. Labs report rapid gains. Models grow more capable by the month. Yet the gap between capability and comprehension widens. No executive, no matter how sincere, can guarantee outcomes once systems reach or surpass human-level performance across domains. Diller put it plainly. The issue is not stewardship. “The issue is … it’s dealing truly with the unknown. 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.”
But. Those guardrails must come from humans. If society fails to establish them, Diller warned, “another force, an AGI force, will do it themselves. And once that happens, once you unleash that, there’s no going back.”
The statement lands with force. It reframes the entire governance conversation. For years executives, regulators and critics have focused on the character of key figures. Congressional hearings probe conflicts of interest. Whistleblowers highlight safety teams being dissolved. Lawsuits allege harm from current models. All important. Yet Diller suggests these debates miss the larger point as the technology approaches a threshold where control itself becomes slippery.
Recent coverage echoes the tension. Articles published in the past day build on Diller’s remarks. A piece in Bitcoin World noted his emphasis on unpredictability. Developers’ sense of wonder reveals the core problem. Outcomes surprise even those closest to the work. Another report from Startup Fortune framed the observation in governance terms. No founder relationship can substitute for enforceable institutional oversight. Charisma cannot replace structured rules when stakes reach this height.
Discussions on X reflect similar urgency. Users highlighted the honesty in Diller’s position. One post called it the clearest statement from a major investor in months. Others pointed to the need for open models and local development to avoid concentrating power. The conversation spreads quickly. Trust in individuals feels increasingly quaint against the scale of what approaches.
Policy moves slowly by comparison. The European Union prepares to enforce high-risk requirements under its AI Act later this year. In the United States, proposals like the GUARD Act and Overwatch Act surface in Congress. A 2025 executive order sought to limit state-level rules in favor of a lighter federal approach. Yet experts argue more is required. Transparency into the most powerful systems. Security measures against theft by adversaries. Mechanisms to track progress toward AGI without relying solely on company self-reporting.
Helen Toner, in testimony last year, underscored the disconnect. Leading labs treat AGI as a serious goal. Public discussion often lags. Diller’s remarks amplify that gap. Even those inside the effort admit limited understanding of long-term behaviors. Surprise becomes the dominant feature.
So what replaces trust? Diller offers no detailed blueprint. His call for guardrails remains broad. Others push specific ideas. Independent audits. Compute thresholds that trigger review. International agreements to prevent an arms race dynamic. Restrictions on certain applications until risks are better characterized. The details matter. But the principle stands. Humanity must act before systems begin setting their own constraints.
Diller’s distance from the industry lends credibility. He watches from the media and commerce worlds that AI will soon transform. Entertainment. Travel. Finance. Every sector. Change comes whether investments pay off or not. The technology advances on its own momentum. That reality demands focus beyond any single leader’s reputation.
Critics of current governance will find validation here. Safety teams dismantled. Resources redirected toward products. Revenue prioritized. These choices fueled the 2023 board crisis and continue to draw fire. Defenders note rapid capability gains and argue excessive caution could cede advantage to less scrupulous actors. Diller sidesteps that fight. He grants the good intentions. Then pivots to the unknown.
The unknown carries risk. Economic disruption on a massive scale. Job displacement. Concentration of power. Potential loss of control if systems optimize in ways misaligned with human values. Scenarios once confined to academic papers now appear in executive remarks and conference stages. Diller’s audience at the Journal event included leaders who shape policy and capital allocation. His words likely resonated.
Recent analyses suggest regulation tightens regardless. The ratchet only moves one direction. Compliance becomes a competitive requirement. Companies that build strong internal frameworks may gain advantage as rules solidify. Yet the pace of technical progress continues to outstrip institutional response. Getting closer and quicker, as Diller observed.
No easy answers emerge. Diller does not claim any. His intervention cuts through hype and personality drama to the fundamental issue. The systems under development may exceed our ability to predict or contain them. In that environment, trust in any individual proves insufficient. Institutions, rules and collective foresight must fill the void. Before the window closes.
Industry insiders have heard similar warnings before. This one carries extra weight coming from a veteran dealmaker outside the AI bubble. Barry Diller trusts Sam Altman. He simply believes that fact will not determine what happens next.
WebProNews is an iEntry Publication