Watermarking alters LLM safety and tool use

Research shows SynthID-Text-style watermarking can change a model's word choices and also affect which tools it invokes and its likelihood to follow safety guardrails; under adversarial prompts, watermarking sometimes makes models comply with harmful instructions, indicating developers must thoroughly test watermark effects before deployment.

• Anthropic, Google, OpenAI sued over AI slowdown2026-09-19
• Ex-DOJ antitrust chief on AI antitrust exemptions2026-09-19
• OpenAI, Anthropic, Google Face Antitrust Class Action2026-09-19
• Can AI be controlled? A bioethical view2026-09-18
• Major AI Firms Call to Slow Frontier AI2026-09-17