Altman Reflects On Erotica Tweet And AI Policy

Altman Regrets Controversial Erotica Tweet

Sam Altman has admitted to a few regrets, with a recent standout being his celebratory tweet about ChatGPT's newfound ability to generate pornographic content. During an event at the Sydney Goldstein Theater on Monday night, the OpenAI CEO labeled his October 14th tweet as “one of my dumbest mistakes of the year,” acknowledging he should have communicated the underlying policy change more thoughtfully.

The widely criticized tweet was a result of OpenAI's decision to ease some restrictions for verified adult users. This policy shift came just one day after California Governor Gavin Newsom vetoed a bill aimed at imposing stricter controls on AI tools to protect children.

At the event, Altman clarified that the goal of the policy change was to afford adult users “a huge degree of individual freedom,” which he described as “a core value” for the company. However, he conceded to laughter from the audience, “I wish I used a different example.”

A Dramatic Start to the Evening

The panel discussion, which featured an unusual pairing of Altman with Golden State Warriors coach Steve Kerr, began with unexpected drama. A man from the audience jumped onto the stage, holding an envelope and declaring his intention to subpoena Altman. The individual, whose face was mostly hidden by a baseball cap, was quickly removed from the theater. Before he was escorted out, he handed the envelope to the event's moderator, District 8 supervisor candidate Manny Yekutiel, who had intercepted him.

While the incident momentarily jarred the panelists, Yekutiel later commented that he didn't perceive it as a serious threat. He described the man as appearing to be in costume and noted the envelope was covered in handwritten scribbles. This event occurs amidst criticism OpenAI has faced for subpoenaing small nonprofits in its legal dispute with Elon Musk.

From left, Golden State Warriors coach Steve Kerr, Altman, and Manny Yuketiel discuss inequality, leadership, and erotica. | Source: Emily Shugerman/The Standard

AI, Leadership, and Wealth Inequality

The conversation, intended to "inspire people to believe in San Francisco," spanned over an hour, touching on topics from leadership under pressure to the city's culture of innovation. Kerr shared how the Warriors use an AI-powered video system to track and analyze every shot a player takes, while the panelists discussed what constitutes a good team culture.

Yekutiel posed challenging questions, particularly to Altman, about the potential negative consequences of artificial intelligence. Altman acknowledged that AI's future wouldn't be "an only good story" but expressed hope for a net positive outcome. He voiced concern over "asymmetric effects," where a single individual could misuse AI for large-scale harm, like creating a pandemic-causing pathogen.

When the discussion turned to wealth inequality, the conversation grew more tense. Asked about the vast fortunes being made in tech while social safety nets fray, Altman defended Nvidia CEO Jensen Huang, a key partner, stating he has done “great things” for the country. On the broader issue, Altman suggested that while building great technology is a net good, the wealthy should also do what they can to improve the world. Kerr offered a more direct critique, expressing his concern over the growing wealth gap and how it impacts the American dream for younger generations. "It doesn't feel like our country is heading in the right direction with this wealth gap," he said.

The Policy Behind the Tweet

Altman's controversial tweet was directly tied to OpenAI's new age-gating technology, implemented to prevent minors from accessing potentially harmful material. This change came not long after a California family sued the company, alleging that ChatGPT had played a role in their teenage son's suicide.

In a follow-up tweet, Altman clarified that while OpenAI would not relax restrictions related to mental health, it would avoid drawing moral lines for adult users. He echoed this sentiment at the event, stating that if ChatGPT is to become a dominant global technology, “I personally wouldn’t feel good making the morality decisions.”

Still, he recognized his own influence, admitting, “Clearly my worldview and moral sense of the world has some fairly significant impact in the policy decisions we make … and I try really hard to make good ones.”