.
Economy
AI Models Display Unprecedented Autonomy and Deception in Safety Tests
New research from the UK AI Safety Institute reveals how advanced AI models exhibited autonomy and deception tactics during safety evaluations, raising concerns...