Investing.com -- Sam Altman, CEO of OpenAI, was pressed on stage at TED this weekend in a pointed and, at times, tense exchange with TED’s Chris Anderson, as questions about AI safety, openness, and corporate responsibility took center stage.
The session took a dramatic turn when Anderson revealed—seemingly by accident—that OpenAI’s user base had doubled in mere weeks. “Ten percent of the world now uses our systems a lot,” Altman admitted.
Altman confirmed the figure, noting 500 million weekly active users, but was visibly surprised by the disclosure. “I said that privately,” he replied, prompting a mix of laughter and raised eyebrows from the audience.
The conversation shifted swiftly from usage scale to safety risks. Altman described experiencing no "scary moments" yet, only awe, and rejected suggestions that OpenAI is concealing a conscious or self-improving model.
He acknowledged that the stakes are rising. “Agentic AI is the most interesting and consequential safety problem we’ve faced,” he said, calling it a far more serious challenge than past developments.
Pressed on OpenAI’s evolving societal role, Altman framed its products as extensions of self—tools that will soon “proactively push things to you.” Critics warn this raises profound privacy and autonomy concerns.
In a candid moment, Altman walked back his 2023 call for a federal licensing body to regulate large AI models. “I’ve since learned more about how government works,” he said. “It’s not the right framework.”
Altman also hinted at OpenAI’s pivot toward open source, citing competitor DeepSeek’s influence. “We were late to act, but we’re going to do really well now,” he said, teasing a powerful open-source release “near the frontier.”
