Tožba pravi, da so se velikanji umetne inteligence nezakonito dogovorili o upočasnitvi tehnologije.

A new lawsuit alleges that major AI firms, including Anthropic, OpenAI, SpaceXAI and Google, colluded to slow down their AI development, violating antitrust laws. The case was filed Friday in the U.S. District Court for the Northern District of California, claiming the companies coordinated to reduce the pace of innovation in order to prioritize safety measures. This collaboration allegedly diminished the value consumers receive from paid AI subscriptions such as ChatGPT, Claude, Grok and Gemini. According to the lawsuit, the coordination occurred primarily on September 12, when Anthropic CEO Dario Amodei published an essay advocating for industry-wide cooperation to decelerate AI advancement. In the piece, Amodei warned that autonomous AI systems could potentially take over the internet within six months. He argued for a more cautious approach, suggesting a three-step plan aimed at "pacing the frontier" of AI development. During an interview with CBS News, Amodei expressed concern about the rapid pace of progress, stating, "I don't think I fully just appreciated what it would actually be like when the progress was as fast as it was." Amodei's proposal received backing from key figures in the AI sector, including OpenAI’s Sam Altman, SpaceXAI’s Elon Musk and Google DeepMind’s Demis Hassabis. The lawsuit asserts that these leaders publicly endorsed Amodei’s call for a slower development trajectory. It further claims that this agreement violates antitrust laws by allowing competitors to collectively restrict market growth. Legal representatives for four named plaintiffs, individuals who subscribe to AI services such as ChatGPT, Claude, Grok or Gemini, are pursuing the case on behalf of a proposed national class action involving other paying users. Nick Rowley, the lead attorney for the plaintiffs, stated, "AI will quickly spin out of human control and could kill us all if we allow AI safety and protocol ... to be controlled by private self-serving agreements between the world's most powerful 'for profit' technology companies." Amodei acknowledged potential legal challenges in his essay, suggesting that the U.S. government might need to facilitate these discussions without direct involvement. He emphasized the importance of building AI responsibly, noting, "If we build in the right way, I think the probability of something bad happening is very low." Representatives for Anthropic, OpenAI, Google and SpaceXAI have yet to comment on the allegations. Meanwhile, Greg Allen, founder and CEO of Decision Tree Research, echoed similar concerns, stating that AI development should align with safety advancements to prevent risks associated with rapidly evolving models.