AI researchers warn over race toward self-improving systems

Current and former researchers from leading artificial-intelligence laboratories are warning that safety precautions may not be keeping pace with the push toward increasingly capable, potentially self-improving systems. Reuters reported that testimonials compiled by Palisade Research included concerns from people associated with OpenAI and Google DeepMind about competitive pressure, internal coordination and oversight. The warnings describe researchers' assessments of risk rather than certain predictions about future AI outcomes.