A movement to cancel ChatGPT has exploded since the Pentagon selected OpenAI’s technology for military use last week, but judging by the turnout for a protest in San Francisco, it hasn’t exactly spilled into the streets.
Around 50 “QuitGPT” protesters carrying signs and writing chalk messages gathered Tuesday outside OpenAI’s headquarters, denouncing CEO Sam Altman’s cozy relationship with the Department of War. Journalists, eager for a story, showed up more punctually than the demonstrators for the scheduled start of the protest.
The protesters had a range of views. There was a contingent that embraces the benefits of the technology but believes OpenAI has gone too far by being in cahoots with the military as the U.S. wages war against Iran.
Among them was a 26-year-old Oakland tech worker whose head was covered by a cardboard robot mask with red LED eyes. He declined to share his name.
“I never go to protests. This is new for me,” he said, adding that he uses AI for everything from programming to generating recipes. “We’re not normally political people. We’re techies, you know – we want to build stuff. What Open AI is doing in terms of building legal mass surveillance technology for the government … is frankly, insane.”
Then there were those who fundamentally oppose the rise of AI in all its forms.
“I hate AI, and I hate war,” said Jennifer Keith, a graphic designer. “AI steals my data, repurposes it, and makes it so I can’t make a living from it.”
She lamented that San Francisco, where she’s lived since 1988, has lost its soul, pointing to the AI advertising around the city.
OpenAI employees using the company’s second-floor gym looked down at the demonstration. “OpenAI there is blood on your hands” and “Quit your job” were scrawled in chalk on the sidewalk. The crowd chanted, “One, two, three, four, we don’t want a robot war. Five, six, seven, eight, no AI surveillance state.”
OpenAI’s rival for supremacy, Anthropic, has been riding a wave of popularity since CEO Dario Amodei refused to give the Pentagon unfettered access to its technology, which he fears could be used to surveil Americans or allow offensive weapons to fire without human oversight.
Until then, OpenAI’s ChatGPT was the popular choice over Anthropic’s Claude, but that has changed amid the political debate. Anthropic’s Claude surpassed OpenAI’s ChatGPT in daily U.S. downloads for the first time Saturday, according to data from analytics firm Appfigures.
River Bellamy, a software engineer who lives in Berkeley, said the Pentagon contract dispute has made him more committed to using Claude. Before, he preferred it because he thought it performed better; now he also values Anthropic’s political stance.
“The red lines that Anthropic had seemed very reasonable to me,” Bellamy said. “I think it is good for warfighters to have access to LLMs — I think they do have legitimate uses. I just want there to be safeguards in place to ensure that the technology is not used for illegitimate purposes.”