> ## Content Index
> Fetch the complete content index at: https://adjacent.media/llms.txt
> Use this file to discover other available public pages before exploring further.

# Open AI Models Face Hidden Backdoor Risks
- URL: https://adjacent.media/signals/open-ai-models-face-hidden-backdoor-risks/
- Published: 2026-10-17T16:14:58.000Z
- Updated: 2026-10-17T16:14:58.000Z
- Description: Fine-tuning and merging open-weight models—standard practices for customization—create vulnerability vectors that are nearly impossible to audit or detect after the fact. A seemingly legitimate model adaptation could embed malicious instructions.
- Author: Jonathan Greene
- Tags: #signal, theme-ai, model security, open-weight models, alignment

Source: [Schneier on Security](https://www.schneier.com/blog/archives/2026/10/backdooring-open-weight-ai-models.html?ref=adjacent.media)

Fine-tuning and merging open-weight models—standard practices for customization—create vulnerability vectors that are nearly impossible to audit or detect after the fact. A seemingly legitimate model adaptation could embed malicious instructions. The barrier to inserting a backdoor is low (modifying weights directly), but the barrier to discovering one is high (you'd need to reverse-engineer what the altered weights do). This makes open-weight AI a supply-chain security problem similar to software dependencies but harder to remediate. Organizations using adapted open models can't easily verify they're getting what they think they are, which matters as these models move from research projects into production systems handling real decisions.