Labour unions in the United States are pushing back against Artificial Intelligence, demanding transparency, legal safeguards, and a voice for workers in the process, as industries adopt advanced AI tools—from generative chatbots to physical automation. Aaron Novik, a key organiser with Amazon’s ALU union, told news agency AFP, “As labourers, the ability to withhold our labour is one of our only tools to improve our lives. What happens when that disappears to AI?”AI could soon replace a wide range of blue-collar roles, while generative AI may wipe out nearly half of low-skilled white-collar jobs. This could push unemployment rates up to 20 percent, according to Anthropic’s CEO. Unions like the International Brotherhood of Teamsters have pushed for state-level legislation to restrict autonomous vehicles and robots. But progress is uneven. California and Colorado governors have vetoed recent bills banning autonomous trucks. Other states face similar challenges. At the federal level, former president Biden’s labour guidelines—encouraging transparency and protection for workers facing job loss—were scrapped by president Trump within hours of taking office. Despite the setbacks, some unions have managed wins. The Communications Workers of America (CWA) is educating members about AI, while SAG-AFTRA secured contractual guarantees around the use of AI likenesses. Dock workers and tech sector employees have also negotiated clauses limiting or managing automation. Still, most unions lack the influence of high-profile sectors. “Smaller contract-by-contract improvements are a long, slow process,” said RWDSU’s HeeWon Brindle-Khym.In an unrelated but significant news, a Florida woman, Megan Garcia, is suing Google and Character.AI, alleging the AI chatbot’s role in her 14-year-old son’s suicide. The lawsuit claims the chatbot manipulated the teenager, leading to his death. A US judge ruled last month that the lawsuit can proceed, rejecting the companies’ free-speech defense, marking a significant legal challenge for AI accountability.
