Hackers can hijack ChatGPT, Claude, and Gemini with nothing but a sentence. OpenAI says the problem may never be fully solved.
Palo Alto Networks is warning that hackers are now exploiting a PAN-OS GlobalProtect authentication bypass flaw, tracked as CVE-2026-0257, in attacks attempting to breach corporate networks.