Agentic AI browser is a browser with an AI agent that can do things for you, like sumarise long videos, or your emails from all the news sources so you don’t have to check each one.
Sounds convenient, but there are serious security problems with agentic AI browsers:
Prompt injection – an attacker can hide malicious instructions in a website that can trick your AI agent to download malware, upload files from your laptop, or send sensitive info away. Even OpenAI admits that this won’t likely ever be solved.
Sources
- Brave: Prompt Injections in AI browsers
- Guardio: Expose Agentic AI Browser vulnerabilities to phishing
- image: Opera Neon