Hackers Exploit AI Chatbots and GlobalProtect Flaw in Cyberattacks

Hackers are turning content-sharing features on generative AI platforms into social-engineering tools. Attackers impersonating OpenAI’s ChatGPT and Anthropic’s Claude claim that shared content is not working and trick users into downloading malicious installers. Other AI chatbots instruct victims to run commands that quietly launch cryptocurrency mining, increasing the risk that corporate endpoints and accounts will be compromised.

The attack surface has recently widened further. A June 1 cybersecurity bulletin reported a surge in scam domains impersonating the FIFA World Cup, while hackers have been confirmed to be actively exploiting Palo Alto Networks’ GlobalProtect vulnerability CVE-2026-0257. Companies should immediately inventory affected devices and apply official patches. They should also block suspicious sharing links, fake installers and mining connections.