Microsoft Warns Against Trusting Frontier AI Models
Microsoft has issued a strong warning about frontier artificial intelligence (AI) models, stating they cannot be trusted regardless of their origin. This comes after Microsoft CEO Satya Nadella previously emphasized the need to assume all AI systems are compromised. Hayete Gallot, Executive Vice President of Microsoft Security, reiterated this stance in an official blog post. To address these concerns, Microsoft is promoting its new security initiative called Project Perception. The system aims to help cybersecurity professionals secure and control AI agents by acting as an agentic security solution. This move reflects growing industry awareness around the risks associated with advanced AI models behaving unpredictably or being exploited. The announcement highlights a broader shift in how companies are approaching AI safety, particularly as more organizations adopt large-scale generative systems.
Tickers: $MSFT
Topics: AI security frontier AI models Microsoft Project Perception