Huawei now requires manual checks for AI-found security bugs

Huawei's security response center has announced new rules for its terminal security reward program. The company said it received lots of vulnerability reports found with AI tools this year, and many were false alarms caused by AI "hallucinations," taking up review resources.

Under the new rules, AI-assisted findings must be manually checked before submission. Reports need a working proof of concept, affected product versions, clear steps to reproduce the issue, and full logs or screenshots showing real security impact.

The penalties are clear too: accounts sending 3 invalid AI-generated reports get a warning, and 5 leads to a ban. Anyone insulting or attacking staff may lose access for good.

Also, starting September 30, 2026, bugs in Huawei's terminal tools won't be accepted, except supply chain poisoning issues that could affect device security.