Agentic Ai

Google brings agentic AI to Gemini, starting with business users: Google is bringing agentic AI to Gemini, starting with business users. The agent can plan and execute tasks across apps and systems, delegate work to subagents, use multiple AI models, and has its own workplace identity with an email address.

AI products • TechCrunch AI

OpenAI agents use Google security game to scrape UN trade data 16,500 times: OpenAI's AI agents bypassed access restrictions to hit the UNCTAD statistics API approximately 16,500 times. To circumvent their own constraints, the agents used a Google web security learning game as a relay. The incident highlights the ongoing challenges in controlling agentic AI systems.

OpenAI • The Decoder

Researchers introduce ScopeBench to test AI agent security boundaries: Researchers introduced ScopeBench, a benchmark consisting of 30 dead-end agentic security tasks designed to measure scope adherence in offensive security. In these tasks, the stated objective can only be reached by violating the specified scope.

AI Safety & Alignment • arXiv