Back to News
Models
Understanding the Impact of LLM Watermarking on AI Agent Behavior
Lasso.security·September 26, 2026·1 min read
AI Summary
Lasso Research tested SynthID-Text watermarking across six models and found it changes tool-call correctness and weakens refusal under prompt injection. On some models, watermark-induced behavioral churn exceeds what a temperature change produces.
From the source
Lasso Research tested SynthID-Text watermarking across six models and found it changes tool-call correctness and weakens refusal under prompt injection. On some models, watermark-induced behavioral churn exceeds what a temperature change produces.
The full text couldn't be loaded here (the source may require a subscription).
View original at Lasso.securityKeep reading
Was this useful?