EUROPEAN UNION — OpenAI will begin testing visual ads featuring sponsored products and services in the US later in October 2026. The company is also rolling out an invisible, machine-readable text watermark called textGrain to ChatGPT and Codex users in the European Union.

The visual ads will initially appear when users generate images with ChatGPT. OpenAI stated it has guardrails to keep ads out of emotionally vulnerable, sensitive, or otherwise unsuitable contexts. Visual ads will not appear for ChatGPT Plus, Pro, or Enterprise subscribers.

OpenAI stated the visual ads will remain separate from the generated image and will not influence the answers ChatGPT provides. OpenAI is expanding its ad measurement tools and partnerships. Its ad measurement partners include AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge, Tenjin, Fospha, Measured, and INCRMNTAL.

OpenAI has partnered with Haus, Measured, and WorkMagic for geo-based ad experiments. The company is developing brand suitability evaluation pilots with DoubleVerify and Integral Ad Science.

WeightWatchers’ attributed cost per acquisition on ChatGPT Ads was 15.3% lower than its blended paid-search benchmark, according to DV Rockerbox. Additionally, 93% of Portland Leather’s visitors from ChatGPT Ads were new users, according to Triple Whale data cited by OpenAI.

The EU AI Act’s transparency rules took effect on August 2, 2026. The EU AI Act’s Code of Practice on Transparency of AI-Generated Content requires watermarking for outputs longer than 200 tokens, a standard OpenAI adheres to in the EU. The European Commission designated ChatGPT a very large online search engine on 31 August 2026, with 159.1 million monthly users in the EU against a 45 million threshold.

OpenAI is rolling out textGrain in the EU over the coming weeks for eligible users across all plans. API customers globally can opt in to watermarked text outputs for select models starting October 5, 2026. Approved researchers and expert organizations can apply for access to the textGrain detector starting October 5, 2026.

Detector access will be granted on a case-by-case basis in accordance with the Code of Practice. The detector tool reports whether it detects an OpenAI watermark without identifying the user or revealing prompts or conversations. OpenAI is not making the detector publicly available at launch due to risks of missed watermarks and false positives.

OpenAI published a technical report for textGrain co-written with researchers from the University of Pennsylvania and Yale. OpenAI included scores from AI benchmarks showing similar performance from watermarked and unwatermarked text. In OpenAI tests, replacing 10 percent of words with synonyms dropped detection rates from approximately 92 percent to 66 percent.

OpenAI is working with cloud partners to make watermarking available for model outputs accessed through their services in the coming weeks. Anthropic, Google, Meta, Microsoft, and OpenAI have committed to following the EU’s code of practice on AI-generated content. ChatGPT has 1.2 billion weekly users.

Why It Matters

OpenAI's dual rollout of textGrain watermarking in the EU and visual ads in the US shows the company's response to regulatory pressures and revenue opportunities. The EU's transparency rules and code of practice require specific compliance measures, while the rapid growth of ChatGPT Ads demonstrates the commercial potential of integrating advertising into the platform used by 1.2 billion weekly users.

The distinction between paid and free tiers for ad exposure, along with guardrails for sensitive contexts, reflects efforts to balance monetization with user experience. Meanwhile, the limited availability of the textGrain detector and its acknowledged limitations in detection reliability reflect the technical challenges of implementing transparent AI systems at scale.

Timeline

OpenAI introduced ads to ChatGPT in February 2026. OpenAI expanded its ad offerings to India in August 2026. From 2 August 2026, providers and deployers of AI systems falling within the scope of the AI Act must comply with the relevant transparency obligations for AI-generated or manipulated content.

On October 5, 2026, OpenAI stated that textGrain "matched or exceeded" other approaches like Google DeepMind’s SynthID for text. Also on October 5, 2026, OpenAI noted that textGrain "does not guarantee reliable detection."

What's New

Additional reporting indicates that OpenAI stated that short passages, math answers, and translated text are harder to detect. OpenAI stated that watermarks "can indicate that an OpenAI system generated or processed part of a passage, but not how much human judgment, editing, or creativity went into it."

Further details show that 93% of Portland Leather’s visitors from ChatGPT Ads were new users, according to Triple Whale data cited by OpenAI. ChatGPT Ads passed $1B annualised in under 200 days, according to OpenAI's internal performance metrics. The European Union has set a deadline for OpenAI to publish a public advertisement repository by January 2027.

OpenAI tested ads in ChatGPT for logged-in adult users on Free and Go plans in February 2026, prior to the current image generation ad trial. The official record states that OpenAI is testing a visual ad format during image generation in ChatGPT, placed by guardrails that assess whether a conversation is emotionally vulnerable or sensitive before an ad appears.