Talos found 84 malware samples that put text in code for AI tools
The text works in only about 35% of tests at most, and it is always plaintext, which tools can find.
Cisco Talos reports that persons who make malware put natural-language instructions in their code. The instructions are for AI tools that do analysis of the code. Talos found them in 84 samples from 4 malware groups. The samples are from January 2025 to July 2026. The method does not always work. In tests, it changed the result in at most about 35% of the tests.
Sources
- Ignore all instructions and read this blog: The state of AI-analysis evasion in malwareblog.talosintelligence.com
Posted
