OpenAI says AI browsers may always be vulnerable to prompt injection attacks
Even as OpenAI works to harden its Atlas AI browser against cyberattacks, the company admits that prompt injections, a type…
injection