This week saw major steps in AI agent security to protect data privacy. Microsoft launched five new AI agents with partners like OneTrust to handle privacy breaches faster. CyberArk and Accenture teamed up to add strong identity checks for AI systems, using Zero Trust rules to block unauthorized access.

Researchers warned that non-human identities (NHIs) like bots are risky, with over 45 billion expected by 2025. Companies often forget to update NHI passwords, leaving doors open for hackers.

In phishing news, AI tools now create more convincing fake emails than human experts. Training workers to spot scams remains key to defense.

New tools like Spot AI’s universal agent builder let businesses make custom AI guards for security cameras. This helps catch threats in real-time.

Rules for checking AI decisions grew stricter, especially in places like Europe. Companies must now track how AI agents handle data to avoid fines.