Browser AI faces unique security risks, like prompt injection attacks, where malicious actors might try to trick Claude into unintended actions, such as sharing your bank information or deleting important files. While we鈥檝e implemented protections, they aren鈥檛 foolproof. Attack vectors are constantly evolving and Claude may hallucinate, leading to actions that you did not intend. We鈥檝e shared our testing results, including possible attack scenarios, so you can make informed decisions, and strongly encourage you to read about the risks before using this product. Read blog post.
First seen: 2025-12-20 23:30
Last seen: 2025-12-21 20:32