Share Facebook Twitter LinkedIn Pinterest Email OpenAI employed an AI “red team” to test and strengthen GPT-5.6, specifically focusing on its defenses against prompt injection vulnerabilities.
By Michael Fawn