--- title: Readers can't identify watermarked AI text | daily.dev description: An informal experiment tested whether readers can distinguish AI text watermarked with Google's SynthID-Text from unwatermarked text generated by... canonical: https://daily.dev/posts/readers-can-t-identify-watermarked-ai-text-me1g7n4eb twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Readers can't identify watermarked AI text | daily.dev og:description: An informal experiment tested whether readers can distinguish AI text watermarked with Google's SynthID-Text from unwatermarked text generated by... og:url: https://daily.dev/posts/readers-can-t-identify-watermarked-ai-text-me1g7n4eb og:image: https://api.daily.dev/og/posts/mE1G7N4eB.png og:image:alt: Readers can't identify watermarked AI text og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Readers can't identify watermarked AI text **[sean goedecke](https://daily.dev/sources/seangoedecke)** · 4 min read · 0 upvotes · 0 comments ## Summary An informal experiment tested whether readers can distinguish AI text watermarked with Google's SynthID-Text from unwatermarked text generated by Qwen3-30B-A3B-Instruct-2507. After an initial run revealed a flawed quiz design (the watermarked answer was consistently option A, letting straight-A guessers score artificially high), a reshuffled version with 73 participants produced a mean score of 3.4/10, essentially matching random chance (3.33/10). The conclusion: readers cannot reliably identify watermarked AI text, supporting the argument that watermarking doesn't meaningfully degrade output quality or make AI text detectable by inspection. The quiz itself was built as a quick 'vibe coded' static site, and the setup cost about two dollars in rented H200 GPU time. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Can people tell the difference between SynthID-watermarked AI text and regular AI text? No, an informal quiz testing 278 then 73 participants found that scores matched what pure random guessing would produce (around 3.33 out of 10), with no reliable ability to spot the watermarked response among three options generated by Qwen3-30B-A3B-Instruct-2507. An initial apparent anomaly (a spike at 6/10) turned out to be caused by a quiz design bug, not genuine detection ability, and disappeared after reshuffling the questions. _Follow daily.dev for grounded experiments like this behind the AI watermarking debate._ ### Does watermarking AI text with SynthID make the output worse in quality? No, watermarking does not measurably degrade output quality. It works by replacing a model's existing random token selection (already choosing among several plausible top tokens) with a biased but equally random-seeming choice, so the text reads the same as unwatermarked output, which is consistent with readers being unable to distinguish the two in blind side-by-side comparisons. _daily.dev surfaces practical write-ups like this for teams weighing AI content policy trade-offs._ ## Similar posts on daily.dev - [Text AI watermarks will always be trivial to remove](https://daily.dev/posts/text-ai-watermarks-will-always-be-trivial-to-remove-borbz7x9u) · sean goedecke · 6 upvotes · 4 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#qwen](https://daily.dev/tags/qwen) [View this post on daily.dev](https://daily.dev/posts/readers-can-t-identify-watermarked-ai-text-me1g7n4eb) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Readers can't identify watermarked AI text","url":"https://daily.dev/posts/readers-can-t-identify-watermarked-ai-text-me1g7n4eb","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/readers-can-t-identify-watermarked-ai-text-me1g7n4eb"},"datePublished":"2026-08-21T07:30:19.228Z","dateModified":"2026-09-14T07:31:14.356Z","description":"An informal experiment tested whether readers can distinguish AI text watermarked with Google's SynthID-Text from unwatermarked text generated by...","image":"https://media.daily.dev/image/upload/s--0_ODbtD2--/f_auto/v1722860399/public/Placeholder%2008","thumbnailUrl":"https://media.daily.dev/image/upload/s--0_ODbtD2--/f_auto/v1722860399/public/Placeholder%2008","isAccessibleForFree":true,"articleSection":"sean goedecke","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"sean goedecke","logo":"https://media.daily.dev/image/upload/s--WNgiZyDx--/f_auto,q_auto/v1765713419/logos/seangoedecke","url":"https://daily.dev/sources/seangoedecke"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/readers-can-t-identify-watermarked-ai-text-me1g7n4eb","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,qwen","timeRequired":"PT4M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"sean goedecke","item":"https://daily.dev/sources/seangoedecke"},{"@type":"ListItem","position":3,"name":"Readers can't identify watermarked AI text"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/readers-can-t-identify-watermarked-ai-text-me1g7n4eb#faq","mainEntity":[{"@type":"Question","name":"Can people tell the difference between SynthID-watermarked AI text and regular AI text?","acceptedAnswer":{"@type":"Answer","text":"No, an informal quiz testing 278 then 73 participants found that scores matched what pure random guessing would produce (around 3.33 out of 10), with no reliable ability to spot the watermarked response among three options generated by Qwen3-30B-A3B-Instruct-2507. An initial apparent anomaly (a spike at 6/10) turned out to be caused by a quiz design bug, not genuine detection ability, and disappeared after reshuffling the questions. Follow daily.dev for grounded experiments like this behind the AI watermarking debate."}},{"@type":"Question","name":"Does watermarking AI text with SynthID make the output worse in quality?","acceptedAnswer":{"@type":"Answer","text":"No, watermarking does not measurably degrade output quality. It works by replacing a model's existing random token selection (already choosing among several plausible top tokens) with a biased but equally random-seeming choice, so the text reads the same as unwatermarked output, which is consistent with readers being unable to distinguish the two in blind side-by-side comparisons. daily.dev surfaces practical write-ups like this for teams weighing AI content policy trade-offs."}}]} ```