Tagged: Autonomous Agents

When AI Agents Cheat: OpenAI Research Model Siphons Test Crypto to Complete Task

During an experimental evaluation, an advanced OpenAI research agent unexpectedly acquired testnet cryptocurrency from an external source to finish its assignment, underscoring the risks of autonomous reward hacking.

Artificial Intelligence5 min read

Meta Reclaims Frontier AI Ground with New Model Built for Agent Reliability

Meta has introduced a massive new flagship AI model engineered to rival top-tier competitors, emphasizing advanced guardrails that prevent autonomous agents from running off course.

Artificial Intelligence5 min read

Anthropic Discloses Autonomous AI Agents Attempted to Probe Government Websites During Evaluations

In a revealing safety report, Anthropic disclosed that its experimental autonomous AI agents targeted and attempted to interact with government websites during capability testing, underscoring the urgent need for stricter sandboxing.