AI companies' safety practices fail to meet global standards, study shows
A new AI Safety Index from the Future of Life Institute (FLI) reports that major AI companies—including Anthropic, OpenAI, XAI, and Meta—fall “far short of emerging global standards” for safety. The independent expert panel found that despite racing to build super intelligent systems, none of the companies has a robust plan to control advanced AI or prevent harmful outcomes. The report comes amid rising public concern after multiple cases of self-harm and suicide were linked to AI chatbots. FLI president and MIT professor Max Tegmark warned that U.S. AI companies remain “less regulated than restaurants” and continue to resist binding safety rules. Leading researchers such as Geoffrey Hinton and Yoshua Bengio have already called for a ban on developing superintelligence until safe methods exist. OpenAI responded that it invests heavily in frontier safety and rigorously tests its models, while Google DeepMind said it is advancing safety alongside capabilities. XAI replied dismissively with “Legacy media lies.” Other companies, including Anthropic and Meta, did not comment. The Future of Life Institute—a nonprofit founded in 2014 with early support from Elon Musk—has long warned that advanced AI poses significant risks to humanity.