A full-fledged courtroom drama – A gebel-wielding judge and polite ju-san unfolded on Monday at King's Cross, London, just a step away from Google Deepmind's headquarters.
Google was on trial for allegations that it would break its AI safety promise.
Participants in the fake production were protesters from Pauseai, an activist group concerned that tech companies are competing for AI with little safety in mind. On Monday, the group gathered near Kings Cross Station to demand that Google be more transparent about safety checks being performed on the most cutting-edge AI models.
Protesters posed outside Google Deepmind's office. Hullangley/Business Insider
Pauseai claims that Google broke its promise at the 2024 AI Safety Summit in Seoul, South Korea when it agreed to consider an external assessment of the model and publish details on how external parties, including the government, were involved in risk assessments.
When Google launched its latest frontier model, the Gemini 2.5 Pro, in April, it didn't do either. The company said it was because the model was still “experimental.” A few weeks later, they released some safety details with a “model card,” which some experts criticized for being too thin, TechCrunch previously reported. The safety report mentioned third party testers but did not specify who they were.
For Pauseai, this is not enough. More importantly, according to the organization, it's not to let the lapse slip and let Google set precedents.
“If Google lets you escape by breaking that word, we'll send a signal to all other labs that safety promises aren't important and we don't have to maintain our commitment to the public,” Pauseai organization's director Ellahughes was gradually inflated to the crowd to about 60 people.
“AI companies currently have fewer regulations than sandwich shops.”
Protesters will demonstrate at the King's Cross of London. Hullangley/Business Insider
There is a lot to worry about when it comes to AI. Economic turmoil. Job displacement. Misreport. Deep fake. As we know, the annihilation of humanity.
Focusing on specific issues in Google Safety Reports is how Pauseai drives certain achievable short-term changes.
About 30 minutes after the protest, several intrigued passersby joined the cause. After a speech from Hughes, the group proceeded to the Google Deepmind office, where fake court production took place. Some Google employees leaving for the day seemed confused when chants rang out “Stop the race, it's not safe” and “Test, don't guess.”
“AI regulations at the international level are in a very bad place,” Pauseai founder Joep Meindertsma told Business Insider, noting that US Vice President JD Vance warned against AI overregulation in the AI Action Summit.
A fake court trial will take place outside the Google Deepmind office. Hullangley/Business Insider
Monday was the first time Pauseai had gathered on this particular issue, but it's not clear what will happen next. The group is involved with members of the UK Parliament who implement these concerns on the flagpole, but Meindertsma refrains from talking much about how Google is involved in the group and its requests (a Google spokesman did not respond to requests for comment on this story).
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
Meindertsma hopes support will grow and refers to a vote that the public is concerned that AI will move too quickly. The group on Monday consisted of people from a variety of backgrounds, including people working in technology. Meindertsma itself runs a software development company and regularly uses AI tools such as Google and Openai.
“Their tools are very impressive,” he said.
Is there anything to share? Please contact this reporter by email hlangley@businessinsider.com Or a signal at 628-228-1836. Use your personal email address and unprocessed devices. Here's a guide to sharing information safely.
