Larry Summers has held top jobs in academia, government, and finance. But Summers’ role as an adviser and board member at ChatGPT maker OpenAI might have been as consequential as any in his career.
OpenAI’s artificial intelligence breakthroughs have already begun to reshape the global economy, with more disruption sure to come. When Summers joined the company’s board in 2023, however, that outcome was far from assured.
Summers was appointed to the OpenAI board in the aftermath of an attempted palace coup that threatened to destroy the high-flying company with the near ouster of chief executive Sam Altman. And while his role was smaller than those of OpenAI’s full-time executives, he clearly gave it a high priority, devoting one day a week to the company’s affairs, as Summers said in a 2024 podcast interview.
Now, Summers is out at OpenAI — along with many of his other roles. The former Harvard president, treasury secretary, and presidential adviser resigned from the company’s board on Wednesday after the release of email messages he had exchanged with billionaire sex offender Jeffrey Epstein.
During Harvard professor Summers’ two-year tenure on the board of OpenAI, the artificial intelligence company transformed itself from an ambitious startup to become a major driver of the US economy.
The humiliating details of Summers’ relationship with Epstein have soiled the name of a man whose reputation for business competence and credibility helped stabilize OpenAI at a critical time.
In November 2023, about a year after OpenAI stunned the world by releasing ChatGPT, the company board voted to oust CEO Altman. The mutineers alleged a pattern of deceitful and dishonest actions by Altman. They also said he was moving too quickly to release AI technologies for public use without ensuring their safety. The board wasn’t even warned in advance when ChatGPT was unveiled in November of 2022.
But their attempt was a dismal failure. Hundreds of OpenAI employees vowed to quit the company if Altman wasn’t reinstated. Within a week, the rebellious members were kicked out, and the Altman faction restocked the board with allies, including Summers.
The blue-chip economist wasn’t a techie, but he brought political and academic gravitas to the board. In addition, Summers was a true believer in the world-changing power of AI. “This could be the most important general purpose technology since the wheel or fire,” he said on Bloomberg TV, a month after the release of ChatGPT. But Summers had little to say about the potential hazards of the technology, and unlike the ousted board members he expressed no interest in slowing the development of more powerful AI systems to ensure their safety.
During Summers’ tenure, OpenAI doubled and tripled down on Altman’s strategy of massive infrastructure investment, to ensure the company’s long-term dominance.
The company expects revenue of $13 billion this year, but it won’t turn a profit. Instead it has committed to investing hundreds of billions in new data centers and cloud computing capacity. For instance, OpenAI has announced a deal to spend $500 billion on a massive Texas data center in cooperation with tech giants Microsoft, Oracle, and Nvidia.
The company will probably need the horsepower. Each AI query uses more computing cycles than an old-fashioned Google search, and up to 10 times as much electricity. This means that OpenAI and competitors like Anthropic, X, and other companies must also boost their investments in order to keep up.
As a result, investment in AI is now one of the biggest sources of US economic growth. One estimate from JP Morgan concludes that AI capital spending in the first half of 2025 played a bigger role in the growth of the US economy than consumer spending.
But OpenAI also faces pushback on several fronts. There are multiple lawsuits from book publishers, authors, and news organizations arguing that their publications were used to train the company’s AI systems without notice and without financial compensation.
OpenAI also faces a lawsuit filed last year by X owner Elon Musk. The suit alleges that OpenAI is conspiring with Apple to monopolize the delivery of AI services on smartphones, at the expense of Musk’s rival AI service, Grok.
And several lawsuits allege that ChatGPT sometimes advises mentally unstable people to harm themselves. The family of one California teenager said the chatbot gave their son step-by-step instructions on how to end his own life. The teenager later died by suicide.
Hiawatha Bray can be reached at hiawatha.bray@globe.com. Follow him @GlobeTechLab.