Wiser Human research

Testing the limits of controlling Agentic AI systems built on Frontier models

Wiser Human investigates the limits of controlling AI agents built on Frontier models, revealing where safeguards and controls may fail as we approach AGI. Through empirical research, we demonstrate real-world risks posed by agentic AI, challenging assumptions about safety. Our goal is to provide compelling evidence that drives necessary changes in AI governance, deployment practices, and risk mitigation strategies.