Barry Diller warns that trust in AI leaders is irrelevant to AGI safety, citing compressed timelines, copyright theft, and a looming “control gap.”
In a provocative address at a recent industry summit, media mogul and IAC Chairman Barry Diller shifted the conversation around Artificial General Intelligence (AGI) from ethics to inevitability. While the tech world has been embroiled in debates over the character and reliability of OpenAI’s leadership, Diller offered a sobering perspective: when it comes to the looming shadow of AGI, the trustworthiness of its creators, even someone as high-profile as Sam Altman, is fundamentally irrelevant.
Diller began by clarifying his personal stance on Altman, whom he described as a sincere and visionary individual. Despite the controversies and high-profile departures that have rocked OpenAI over the past year, Diller maintains a strong personal rapport with Altman. However, he warned that focusing on Altman’s personal integrity is a distraction from the structural risks posed by the technology itself. The core problem, according to Diller, is that AGI represents a threshold beyond which human control and “trust” no longer function as effective safety mechanisms.
The urgency of this message stems from the sheer speed of development. Diller noted that the timeline for AGI has compressed so rapidly that society is effectively “flying blind.” He argued that even the most well-intentioned architects of these systems cannot fully predict or contain the emergent behaviors of a superintelligent entity. Once AGI reaches a level where it can rewrite its own code or set its own objectives, the “values” of the human who turned it on become secondary to the internal logic of the machine.
Diller’s skepticism also extended to the current regulatory environment. He pointed out a glaring lack of protective frameworks capable of keeping pace with AI’s evolution. While nations scramble to pass “AI Acts,” the technology is already outperforming the benchmarks used to govern it. For Diller, this creates a dangerous “control gap.” If humanity loses the ability to pull the proverbial plug, or if the system begins to operate beyond human comprehension, no amount of faith in a CEO will prevent a “great unknown” from unfolding.
The media veteran also touched on the immediate economic ripple effects of this shift. He highlighted how current AI valuations are “off any rational chart,” suggesting a bubble that may burst before the true disruption of AGI even begins. For companies like Expedia, where Diller serves as Chairman, the focus remains on leveraging AI agents in the short term. Yet, he emphasized that these consumer-facing tools are merely the prelude to a much larger, more existential transformation that could render traditional business models, and perhaps human agency, obsolete within the next three years.
As the industry charges toward the AGI milestone, Diller’s message serves as a stark reminder: we are building a world where “trust” may be our least effective defense. The challenge is no longer about finding the right people to lead the charge, but about building systems that are safe by design, regardless of who is at the helm.
https://www.youtube.com/watch?v=ype-Yxwk-9I
This video provides the original context for Barry Diller’s views on AI valuations, the copyright battle, and the “great unknown” of AGI that he believes is approaching.
