Sam Altman’s reputation gets dragged further into the mud

OpenAI CEO Sam Altman speaks to members of the media during the OpenAI DevDay event on November 06, 2023 in San Francisco, California.
OpenAI CEO Sam Altman.
Justin Sullivan—Getty Images

OpenAI CEO Sam Altman and his husband Oliver Mulherin have just joined the likes of Bill Gates and Warren Buffett in pledging to give away half their fortune, which is nice, because Altman’s reputation is otherwise taking a real battering right now.

Remember the chaos last November when OpenAI’s board abruptly fired Altman, and he got hired by Microsoft for the blink of an eye before returning to OpenAI, as the board members behind his brief defenestration had to leave the building instead? At the time, the lack of clear reasoning behind the ouster—Altman had been “not consistently candid in his communications with the board”—painted a picture where Altman has been wronged by a bunch of clumsy amateurs for no good reason.

But now we know a lot more about what went down, thanks to ex-board member Helen Toner and her appearance on the Ted AI Show podcast yesterday. Here are Toner’s most eye-catching allegations:

—OpenAI announced ChatGPT in 2022 without giving its own board any warning.

—Altman didn’t tell the board he owned OpenAI’s Startup Fund (over which he relinquished control last month,) while claiming to have no financial interest in OpenAI.

—“On multiple occasions, he gave us inaccurate information about the formal safety processes that the company did have in place, meaning that it was basically impossible for the board to known how well those safety processes were working or what might need to change.”

That is, indeed, a lack of consistent candidness. Or, put another way, it seems that Altman did his damnedest to stop OpenAI’s board from exercising its mission of ensuring safe AI development. If Toner’s allegations are true, then I’m not at all surprised that the board fired Altman.

OpenAI said in a statement that it was “disappointed that Miss Toner continues to revisit these issues,” and pointed out that the company’s post-fiasco independent review “concluded that the prior board’s decision was not based on concerns regarding product safety or security, the pace of development, OpenAI’s finances, or its statements to investors, customers, or business partners.”

Altman certainly emerged from the November episode as the winner, with his enemies vanquished, and there is no question that the man knows how to drive business. Just today, OpenAI scored consultancy giant PwC as both ChatGPT’s biggest customer, with 100,000 users, and its first enterprise reseller.

But the aphorism that someone is “only as good as their word” is still something that people take seriously, and for good reason. Coming so hot on the heels of the Scarlett Johansson AI voice drama, the exodus of OpenAI employees who say the company isn’t taking safety seriously enough, and the revelation that OpenAI never followed through on its promise to give proper resources to a key safety team, Toner’s allegations provide yet another reason to view Altman as fundamentally untrustworthy.

As my colleague Jeremy Kahn noted in yesterday’s Eye on AI newsletter, Altman’s lack of probity and the inability of OpenAI’s board to properly oversee him have (together with other things) major implications for AI regulation: “We desperately need a regulator with enough expertise and authority to look over the shoulder of these companies and ensure they aren’t building systems that pose extreme risks.”

Jeremy also recently pointed out that Microsoft CEO Satya Nadella, whose company continues to be OpenAI’s biggest partner, can’t be happy with the amount of ethical unease surrounding Altman and his operation. After what Toner just revealed about how Altman deals with his supposed overseers, I’ll bet Nadella is thanking his lucky stars that he’s not Altman’s boss.

More news below.

David Meyer

Want to send thoughts or suggestions to Data Sheet? Drop a line here.

NEWSWORTHY

TikTok offer revelation. In 2022, TikTok offered the U.S. government a deal that would have allowed officials to hand-pick its U.S. operation’s directors and veto each new hire in the country, the Washington Post reports. Chinese-owned TikTok would also have let a Defense Department partner monitor its source code, and would even have given the government a full-on kill switch for TikTok in the U.S. But the Biden administration refused, and TikTok now must find a U.S. buyer or be banned from app stores. (TikTok’s lawsuit over that new law will proceed to oral arguments in September.)

Microsoft-Tencent deal. Microsoft will make Tencent’s Android apps playable on Windows PCs, under the terms of a new agreement that will see around 1,500 such mobile apps listed in the Microsoft Store in China. As The Register reports, Microsoft previously had a similar Android-Windows deal with Amazon, which lasted less than two years. In other Tencent news, the company’s Dungeon & Fighter Mobile game has provided its first big hit in a while, generating an estimated $140 million in player spending during its first week of release in China.

Google Search leak. There appears to have been a huge leak of documentation from Google’s Search division, giving search-engine optimization experts an unusual insight into how Google ranks its results, and seemingly contradicting statements that Googlers have made on the topic. As The Verge notes, Google has yet to comment on the leak.

SIGNIFICANT FIGURES

6%

—The rally in Nvidia’s share price yesterday, bringing the chipmaker’s market value ($2.8 trillion) within striking distance of Apple’s ($2.9 trillion).

IN CASE YOU MISSED IT

How Waymo outlasted the competition and made robo-taxis a real business, by David Meyer

Malaysia’s prime minister touts the country as a ‘neutral and non-aligned’ home for chip companies amid U.S.-China tensions, by Lionel Lim

Generative AI is in the crosshairs at Meta and Alphabet annual meetings as shareholders vote for detailed reports, by Sharon Goldman

Deflated ‘unicorn’ CEO Adam Neumann surrenders after judge kills his attempt to buy back WeWork, by Dylan Sloan

GameStop surges without the help of Roaring Kitty, by Chris Morris

Hacktivist attacks on Europe have doubled since 2023, top EU cybersecurity official says: ‘This is part of the Russian war of aggression’, by the Associated Press

BEFORE YOU GO

Russian software fees. Russian companies already have a tough time getting their hands on (legally acquired) foreign enterprise software due to sanctions, but many still use it, and the Kremlin would like to change that. According to Reuters, Vladimir Putin’s government is considering a new levy on foreign software, to “equalize” the playing field with Russian software firms. “We believe that we need to create additional economic incentives for businesses to switch to Russian solutions and this should be linked to certain tax motivations,” Digital Development Minister Maksut Shadaev reportedly said yesterday.

This is the web version of Fortune Tech, a daily newsletter breaking down the biggest players and stories shaping the future. Sign up to get it delivered free to your inbox.