+1 on better governance/oversight +1 on provenance and watermarking of AI-generated works +1 on funding for AI safety research +1 on funding for economic/societal impact research I'd like to see more transparency In terms of the models and the training data - the idea that nobody understands the emergent capabilities and can look inside the black box is a big worry. I'd also like to ensure that content creators/owners can opt out of model training through some enforceable CC-like license meta-tagging. Finally - I worry that people will hijack this open letter to simply slow down the competition or, in the case of Elon Musk - scuttle OpenAI because he feels he's been wronged. #ai #safety https://futureoflife.org/open-letter/pause-giant-ai-experiments/