A viral post on X claims that a new OpenAI model identified 35 vulnerabilities in Apple devices without naming the model or detailing the flaws. However, Apple’s security notes credit Codex Security with only three WebKit vulnerabilities, not 35, although the June 2026 patch batch addressed more than 30 flaws with assistance from artificial intelligence. In November 2026, OpenAI announced that an autonomous cybersecurity pipeline had discovered 948 distinct local privilege-escalation vulnerabilities in the Linux kernel and then generated exploits and patches. The campaign, conducted in October, reportedly used approximately $1.2 million in API credits. No major security incident has yet been linked to these findings, but Apple has accelerated its patch-release schedule in response to a threat cycle that artificial intelligence is helping accelerate.
Source: Read the original article


