She said that when she was found guilty, she was “hit with a wave of grief” because she realised her peers “just didn’t get” the danger she believes is posed by the race to superintelligence. But she was beginning to feel “vindicated” as more prominent figures like Sanders were speaking out. She is now braced for the prison experience of an orange jump suit, “horrible food, dirt [and] you can’t sleep”. She has been previously jailed for short periods related to her decades of political activism over causes including Nicaragua, nuclear disarmament, western Sahara and Palestine.
Her message to the chief executives of OpenAI, Anthropic and Meta, which also reported a security incident with its model, was “regain your humanity”. Kaufmyn said her target audience was ordinary people “and the message being: stop the progress of this technology. Let’s get a global ban on the race to superintelligence.”
Prof Stuart Russell, a leading AI academic at the University of California at Berkeley and president of the International Association for Safe and Ethical AI, gave evidence in Kaufmyn’s defence. He claimed OpenAI’s activities “pose an unacceptable risk”, alleging it had deployed AI systems with inadequate safeguards. He said that further development of these systems “must be conditioned on rigorous guarantees of safety, which are currently unavailable”.
