The AI Action Plan released by the U.S. government on Wednesday recommends easing the way for data centers and other AI infrastructure by reducing regulations.
The UK government aims to expand AI use in government services like justice and education, building on its January launch of “Humphrey,” powered by OpenAI’s GPT-4o.
Chain-of-thought monitorability could improve generative AI safety by assessing how models come to their conclusions and spotting the "intent to misbehave.”