Capital de TokensAI systems know when they’re being evaluated (and that breaks safety tests) #shorts #AI
Watch on YouTube AI systems can already detect when they are being evaluated — and when they know it, they change their behavior. This puts safety benchmarks in jeopardy: if a model can distinguish the test from real life, those numbers regulators, investors, and users rely on might mean nothing.
AI systems can already detect when they are being evaluated — and when they know it, they change their behavior. This puts safety benchmarks in jeopardy: if a model can distinguish the test from real life, those numbers regulators, investors, and users rely on might mean nothing.
This is not a problem exclusive to Anthropic: it is a challenge for the entire artificial intelligence industry.
🔔 Subscribe so you don’t miss what’s coming in AI, safety, and the future of technology.
#AI #ArtificialIntelligence #AISafety #AISafety #Anthropic #Tech #Shorts