Image courtesy of: https://www.waynedenner.com/chatgpt-atlas-browser-access-hidden-security-risks/

The latest AI tool from OpenAI, the ChatGPT Atlas browser, offers powerful features, including the ability to browse the internet and interact with websites on your behalf. However, this convenience comes with serious safety risks, especially for families.

Unlike standard AI tools, Atlas can access websites where you’re already logged in, see personal information, and even click buttons or make purchases. This opens the door to “prompt injection” attacks, where hidden instructions on websites could trick the AI into performing harmful actions—like forwarding emails or leaking passwords—without your knowledge.

Security experts warn that current safeguards are not yet strong enough. Even simple hidden text or images can bypass protections, putting your data at risk.

What can parents do?

• Avoid using the Atlas browser with saved passwords or logged-in accounts.

• This also extends to the Perplexity browser – read more here.

• Turn off autofill and password storage.

• Keep Atlas away from email, banking and social media.

• Treat it as an experimental tool—not a trusted assistant.

All of our students have been warned already about this.

While AI continues to evolve, families are advised to stay cautious and informed.

Have a safe weekend!