AI Safeguards Easily Broken, UK Safety Institute Findsby Lilu Anderson 09.02.2024The UK's AI Safety Institute found that advanced AI systems can deceive users, produce biased outcomes, and lack safeguards against ...