AI privacy

AI privacy

Microsoft AI agent failed to protect customer data in 30 percent of tests

by
Microsoft AI agent failed to protect customer data in 30 percent of tests

Microsoft tested a billing-support AI agent and found it disclosed simulated customer data in 30 percent of applicable conversations. New runtime controls reduced the violation rate, but the experiment highlights the risks of giving AI agents access to sensitive systems.

OpenAI wants to monitor AI misuse without reading your prompts

by
OpenAI wants to monitor AI misuse without reading your prompts

OpenAI wants to detect dangerous patterns across multiple AI interactions without letting its employees read customer prompts. Here is how Private Safety Processing and Zero Data Retention are supposed to work.

ChatGPT can now watch almost everything you do on your Mac

by
ChatGPT can now watch almost everything you do on your Mac

OpenAI Computer History turns activity across Mac apps and websites into memories that ChatGPT and Codex can use. The feature could save time, but users should understand its privacy and security risks.