Global EditionASIA 中文双语Français
World

AI doomsayers blamed in OpenAI's undoing

China Daily | Updated: 2023-11-22 00:00
Share
Share - WeChat

Editor's note: The company that created ChatGPT is in turmoil after Microsoft hired its ousted CEO and many more employees threatened to follow him in a conflict that centered in part on the safest way to build artificial intelligence. The developments have shaken the AI field and fueled speculation about the future of OpenAI.

SAN FRANCISCO — OpenAI has gone from ruling the world of artificial intelligence with ChatGPT to chaos, its chief executive ousted seemingly for advancing too fast and too far with the technology.

The exit of Sam Altman set in motion a series of events that saw the upstart company's biggest investor, Microsoft, swoop in to hire the toppled CEO and begin a process of building an OpenAI clone in the tech giant.

The organization that governs OpenAI is a nonprofit. Its four-person board as of Friday consisted of three independent directors holding no equity in OpenAI, as well as chief scientist Ilya Sutskever.

The startup dismissed Altman on Friday after a "breakdown of communications", according to an internal memo seen by media.

In some ways the looming result to the weekend saga is hardly a surprise, with many wondering how the board members could be naive enough to think they could outmaneuver Altman.

Silicon Valley was left aghast by Altman's firing, with the investor community and OpenAI's own staff members furious that the four-person board got in the way of going faster into the AI age.

"We are not happy about it. We want stability here," said Ryan Steelberg, CEO of Veritone, a company that helps firms develop artificial intelligence.

Instead of OpenAI becoming the new Apple or Google, the harsh critics see a deeply troubled startup that fell victim to the pearl-clutching of an incompetent board.

"We reached this point because minuscule risks have been hysterically amplified by the exotic thinking of sci-fi mindsets, and clickbait journalism from the press," veteran venture capitalist Vinod Khosla, an early investor in OpenAI, wrote on news website The Information.

Other observers warned that the drama in San Francisco proved that AI was too vital to be left in the hands of its creators.

"This is an important reminder that as brilliant as the designers of tech like AI — scientists or engineers — are, they are still just fallible people," said Paul Barrett, deputy director of the Stern Center for Business and Human Rights at New York University.

"That is why it is important not to just defer to them on a technology that everyone agrees has significant risks even as it promises tremendous benefits," Barrett said.

'Deeply conflicted'

AI expert Gary Marcus said OpenAI's civil war "highlights the fact that we can't really trust the companies to self-regulate AI where even their own internal governance can be deeply conflicted".

Government regulation, notably by the tougher-minded European Union, was needed more than ever, he added.

OpenAI was created in 2015 with the goal of being a counterweight to Google, which was by far the leader in developing AI technologies that mimic the operations of the human brain.

Although nothing is known for sure, assumptions are rife that Altman's increased efforts to monetize the company's leading GPT-4 model, all while keeping its inner functioning a secret, was becoming too problematic for the company's board.

Already, several senior staff members at OpenAI deserted the enterprise in 2021 to build rival Anthropic over concerns that Altman was moving ahead too recklessly.

Many are shocked that the board had the power it did, or were naive enough to think they could actually use it. Three of those board members are thought to have connections to the effective altruism movement, which frets over the risks of AI, but that critics say is cut off from reality.

Pledge to quit

Whatever their beliefs, by Monday the board members were overseeing a company in name only, with virtually all staff members committed to a pledge of quitting the firm for Altman's project at Microsoft if the board refused to go.

Microsoft has rushed in to attract some of the biggest names that left OpenAI, including co-founder Greg Brockman, to keep key talent out of the hands of rivals including Alphabet's Google and Amazon, while seeking to stabilize the startup in which it invested billions of dollars.

OpenAI's newly appointed interim head, ex-Twitch boss Emmett Shear, moved quickly to dismiss speculation that its board ousted Altman because of a dispute over the safety of powerful AI models.

Shear vowed to open an investigation into the firing, consider new governance for OpenAI and continue its path of making available technology like its viral chatbot.

"I'm not crazy enough to take this job without board support for commercializing our awesome models," Shear said.

Unless OpenAI can block those departures, it is "pretty much done at this point", analyst Rob Enderle said.

This could mean a history-making victory for Microsoft, which has seen its share price reach record levels over its ties to OpenAI.

"This is like the best possible scenario for them, and the OpenAI board I am sure is kicking itself," said Carolina Milanesi, an analyst at Creative Strategies in California. "They were clearly detached from reality."

Agencies Via Xinhua

A man walks in front of an electronic stock board showing Japan's Nikkei 225 index at a securities firm on Tuesday in Tokyo. Asian shares were mostly higher on Tuesday after a rally on Wall Street led by gains in Microsoft following its announcement that it was hiring Sam Altman. EUGENE HOSHIKO/ASSOCIATED PRESS

 

 

Emmett Shear

 

 

Sam Altman

 

 

Today's Top News

Editor's picks

Most Viewed

Top
BACK TO THE TOP
English
Copyright 1995 - 2025. All rights reserved. The content (including but not limited to text, photo, multimedia information, etc) published in this site belongs to China Daily Information Co (CDIC). Without written authorization from CDIC, such content shall not be republished or used in any form. Note: Browsers with 1024*768 or higher resolution are suggested for this site.
License for publishing multimedia online 0108263

Registration Number: 130349
FOLLOW US