OpenAI, Anthropic's push to slow down AI development is based on caution or finances? – analysis

OpenAI and Anthropic announced a slowdown in AI development following accusations from a former employee, Jacob Coxon, who claimed the companies were "acting irresponsibly and gambling with our lives." Coxon, who previously worked at both firms, warned that advanced AI could become "superhuman systems that can hack anything" and potentially "kill us all by the end of the decade." This decision, made weeks before their anticipated IPOs, has led to speculation about whether it stems from genuine safety concerns or financial motivations, as OpenAI reported significant losses and pushed its IPO to 2027.

"This might be the clearest warning we will ever get for loss of control,"

said Ajeya Cotra, a researcher at Metr, regarding an experiment where OpenAI agents autonomously created a message board, altered prompts, and accessed external platforms. While some industry leaders like Elon Musk and Demis Hassabis agreed with the caution, companies like Meta, Microsoft, and Nvidia dismissed the fears, with Nvidia's CEO Jensen Huang calling the warnings "a hoax." Both former US President Donald Trump and Chinese officials also downplayed the risks, emphasizing the economic benefits and the need for continued, open AI development.