OpenAI makes textGrain watermarking opt-in for API customers everywhere

Article 50 of the EU AI Act requires providers of generative AI to mark their output in a machine-readable way, and it has applied since August 2, 2026 ^[14]. Crypto Briefing says it covers anyone serving EU users ^[15]. City A.M. notes an exception for AI that only assists with editing without substantially changing the original meaning ^[16]. Whether a company calling OpenAI's API counts as a provider is a legal question. Tech AI Wire's advice to teams serving EU users is to ask their legal department ^[18].

On the API side, the switch is set per project, according to Unite.AI as cited by Tech AI Wire ^[3]. Customers on Microsoft Azure get the same option through OpenAI's cloud partners in the coming weeks ^[5]. The reports do not include the parameter name or say whether a single request can override it.

textGrain biases which words the model picks, and the detector tests a passage for that statistical pattern ^[6]. Length and subject decide the result ^[7][8]. At a 1% false-positive target, OpenAI reported about 95% detection on 400-token psychology passages and about 80% at 200 tokens ^[7]. At 200 tokens, roughly one marked passage in five goes undetected ^[26]. Maths output scored "substantially lower" because the model has less freedom to choose words ^[8].

For the 95% figure to transfer to a product, its output has to resemble OpenAI's test: around 400 tokens of prose with room in word choice, left unedited. Editing is where it breaks ^[10]. Swapping 10% of words for synonyms cut detection on 400-token passages from about 92% to 66%, and swapping a quarter cut it to 17% ^[10]. Replacing a quarter of the words removes 75 points of detection ^[27]. Robustness has stalled OpenAI before: Crypto Briefing reports that a 2024 prototype was "about 99.9% effective" and was held back over concerns about how robust it was ^[20].

A 1% false-positive rate means about one human-written passage in 100 is wrongly flagged ^[21]. Across 10,000 human passages, expect about 100 false flags ^[28]. OpenAI also says a missing watermark does not prove a human wrote the text ^[22]. Few teams can run the check yet: the detector is limited to selected researchers and specialist organisations, who apply through a form and are approved case by case under the EU Code of Practice ^[12]. OpenAI said it will widen access "when we believe results can be interpreted responsibly," and has not said when ^[13].

On quality, OpenAI tested its frontier model Astra on eight benchmarks, including GPQA Diamond, BrowseComp and DeepSWE, and found no significant difference with the watermark on or off ^[9]. A good DeepSWE score says little about whether a cover letter still sounds like its author ^[9]. The Decoder notes that the results do not establish whether watermarking affects writing quality, a point critics have also raised about Claude's watermark ^[11].

Teams on Claude skip the decision. Anthropic's watermark is mandatory and applies globally however the model is reached ^[4]. City A.M. says the marks survive copy and paste, and that human-written text edited by Claude may carry them ^[24]. Anthropic says its system holds up well against edits but has not published detection rates ^[23]. I think OpenAI's disclosure is the better engineering practice: a technical report, a stated false-positive target, failure rates by length, subject and edit fraction, and a plan to open-source the method ^[19][7][10]. OpenAI also says textGrain matched or beat Google's SynthID for text in internal tests, and SynthID is the technology The Decoder says Claude's watermark uses ^[25][17].