OpenAI Under Pressure: Protests Against Military AI

In San Francisco, in front of OpenAI's headquarters, about thirty protesters chanted slogans such as "1, 2, 3, 4, we don't want a robot war!" and "5, 6, 7, 8, no AI surveillance state!" This gathering aimed to raise awareness about the dangers of a surveillance state powered by artificial intelligence. Although the number of protesters was small, they managed to draw attention to a hot topic.

While the protesters voiced their concerns, a small group, likely OpenAI employees, entered and exited the building without making eye contact. This scene sharply contrasted with the dark tone of the slogans, creating an atmosphere of dark comedy.

The main reason for this protest is the replacement of Anthropic by OpenAI as the primary AI service provider for the Pentagon. This decision was announced last Thursday when the Pentagon labeled Anthropic as a "supply chain risk," a designation typically reserved for foreign adversaries.

OpenAI is now negotiating with the Pentagon to strengthen the guarantees of its contract, after a previous attempt failed to reassure the public. Anthropic is also continuing negotiations with the military, despite being sidelined.

This protest is part of a broader wave of anti-AI demonstrations that have taken place across the United States. Last week, hundreds of people marched in front of the headquarters of DeepMind, OpenAI, and Meta during a protest that, according to organizers, attracted up to 500 participants. The previous month, 200 protesters marched on the Virginia state capital to oppose the construction of data centers in the region.

For OpenAI, initial doubts about the robustness of its guarantees against the use of its technologies for domestic surveillance and autonomous weapons have turned into a full-blown public relations crisis. On Tuesday, the company's negotiations led to an impromptu protest outside its headquarters in San Francisco, attracting around thirty people. OpenAI did not respond to a request for comment on the matter.

As I arrived at the protest, I wondered what kind of people would show up and what their demands would be. In the previous week, posts on Reddit regarding ChatGPT opt-outs had received tens of thousands of upvotes. Comments ranged from "You are now training a war machine. Let's see the proof of cancellation" to "I love celebrating any kind of downfall for OpenAI." Discussions revolved around the AI bubble, aversion to Trump, data protection, artist protection, and the water and electricity costs of data centers.

Among the protesters, some could be perceived as anti-AI Luddites, others as hippies concerned about the impact of data centers on local water resources, or even technocrats loyal to Claude. I took out my notebook and pen to talk to them.

One organizer, wearing a cardboard robot mask, introduced himself as "The Last Friendly Robot." He explained that he was there to protest against mass surveillance, which he attributes to the "Epstein class." Although Epstein is not known for advocating surveillance, the robot referred to the trend of billionaires developing surveillance technologies while preserving their own privacy.

The robot expressed particular concern about AI-powered weapons, capable of "killing without conscience, and without a human in the loop." Having worked in technology, he knew people in major AI companies and believed that OpenAI's new contract betrayed the initial promise of the technology: "This is not what we came to the Valley for. We came to build technology that liberates people!"

This sentiment was shared by other protesters, who viewed OpenAI employees as complicit in something harmful. "Quit your job!" was another protest slogan, chalked on the sidewalk among many colorful drawings made by the protesters.

I also spoke with Perrin Milliken, a young woman working for a climate NGO. She shared her concerns about "data centers, and how they consume so much water and electricity" in local communities. While data centers can sometimes put pressure on water supplies in certain areas, experts point out that reported statistics on AI's water consumption are often exaggerated. However, the issue of electricity is very real, as evidenced by recent commitments from AI labs to offset price increases that would otherwise be borne by consumers.

Milliken had not closely followed the Defense Department issue, but she expressed concern about OpenAI's recent actions. "It was scary to see how quickly they agreed to everything the government wanted," she said. She felt compelled to switch to Gemini from ChatGPT due to the events of the week. "We can't be complacent right now," she added, highlighting a moment filled with "very powerful and wealthy people who do not have the interests of the people at heart."