AI Text Watermarks like SynthID Are Extremely Subtle and Fail on Short Outputs

giffmana · x · 2026-08-12

Discussing AI text watermarking, the author notes that systems like SynthID are designed to be extremely subtle and unnoticeable to users. Furthermore, watermark strength depends on text length; if the model's output is very tiny, there simply isn't a strong watermark present.

Related event: AI Text Watermarking: Mechanisms and Limits, Low-Entropy Outputs Hard to Mark, but Social Benefits Outweigh Costs(6 posts)→

Original post →

More from Models

• First Model to Pass the Video Turing Test: Half of Users Thought It Was Human — Puzzleheaded-King584 · 2026-10-03
• Using System One models in Swift: fast, deterministic decisions via Apple Foundation Models — rxwei · 2026-10-03
• Linux Kernel CVEs Surge From ~500 to 1500+ Per Release, LLMs Blamed for Bulk of the Rise — burny_tech · 2026-10-03
• Developer Complains OpenAI's Coding Model Endlessly Scopes Creeps Instead of Finishing Tasks — DavidWells · 2026-10-03
• Sonnet 5 Spotted in Google Antigravity Backend, Which Still Runs Sonnet 4.6 — brandon_galang · 2026-10-03
• Every's Dev Day chat with Matthew Berman: budget gone the moment he tried Ultrafast — every · 2026-10-03

Models channel →