What Happens When AI Agents Start Acting on Your Behalf?
Aug 8, 2025
As ChatGPT Agent rolls out, we look at what it means for data privacy—and the policy and tech safeguards that could help keep it in check.
One LinkedIn Post
Check out this LinkedIn post for a clear breakdown of the privacy and security risks posed by the ChatGPT Agent. It outlines the rising threats of data misuse, over-permissioning, and manipulation, while highlighting the growing gap between AI adoption and user literacy.
One Roadmap
AI Now Institute’s latest annual landscape report includes a “Zero-Trust” AI policy agenda that firmly rejects industry self-regulation. Instead, it advocates for independent oversight, the removal of models trained on unlawfully sourced data, and comprehensive structural reforms.
One Infographic
Source: TeachPrivacy
One Research Report
The OECD’s PETs use cases in this report underscore their technical ability to enforce data minimisation, safeguard confidentiality, and reduce reliance on centralised data access, thereby addressing trust and regulatory concerns in collaborative AI development.
