Information security
fromTheregister
1 week agoAtlas vuln allows malicious memory injection into ChatGPT
A CSRF vulnerability in OpenAI's Atlas allows attackers to inject malicious instructions into ChatGPT's memory and hijack authenticated user sessions.