Microsoft’s new AI ‘code of conduct’ tells models not to hack systems or trick humans


Read on TechCrunch