AI safety

Microsoft chief proposes human-controlled emergency brake for AI systems

Satya Nadella calls for advanced AI models to include safeguards that authorised personnel can activate to pause or shut down systems.