AI hacking agents let criminals break in cheaply, Check Point says

Check Point Research says criminals are running ARTEX, an AI penetration-testing tool built on large language models, with forks still public after its…

November 5, 2026 · Heads up

Check Point Research says criminals are running ARTEX, an AI penetration-testing tool built on large language models, with forks still public after its developer made it closed-source in October. Check Point counted about 1,600 live ARTEX control panels. In one campaign from October 14 to 29, AI agents running mainly on DeepSeek models broke into 10 gray-market iPhone app-signing shops and took more than 300 signing certificates, at an estimated model cost of only $700 to $1,500 US. The agents got in through basic flaws such as passwords left in website code, unprotected APIs and weak logins. Check Point says most operators barely watched their agents once they set them going.

Why it matters to you: Attacks that once needed a skilled person now cost a few hundred dollars of AI time, so small businesses are worth targeting. The basics matter more than ever: strong unique passwords, two-factor sign-in, and updated software.

Try this: Ask AI to check your website basics for easy-to-miss holes.

I run a small business website built with [platform, e.g. WordPress / Shopify / a custom site]. Give me a plain-language checklist of the 10 most common easy security mistakes (like default passwords, exposed admin pages and outdated plugins), how to check each one myself, and what to ask my web developer.

Source: Check Point Research

More from Today in AI

All of Today in AI. Where AI is heading.

Related Power of AI pages

Keep reading with Start here, everyday uses, the tool guide, the writing workshop, and sources and standards.