OpenAI brings visual ads to ChatGPT image generation this month

The latest:

A new visual ad format is coming to ChatGPT, built around images that show product inspiration, usage or experiences, OpenAI said. Testing begins later this month in the United States with an initial group of advertisers, starting inside image generation. OpenAI said advertising lets it serve a much larger market than subscriptions alone.

Details:

• The format: The ads are visual rather than text-based, designed to help people picture how a product or service could fit into their lives. They will appear first during image generation in ChatGPT, kept separate from the image being created, and clearly labeled as advertising.
• The firewall: OpenAI said advertising does not influence ChatGPT’s answers. The company also said placement guardrails assess whether a given conversation is appropriate for advertising at all, holding ads back from emotionally vulnerable or sensitive contexts rather than treating every chat as inventory.
• The scale: OpenAI put ChatGPT’s reach at 1.2 billion people each week, describing it as the world’s largest AI-native consumer platform. That audience is the core of the commercial pitch: a user base built on subscriptions now being opened to advertisers for the first time at this scale.
• Measurement push: New integrations with Hightouch, Tealium and LiveRamp will let advertisers send conversion data into ChatGPT Ads. OpenAI also added support for attribution partners including AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge and Tenjin.
• Deeper analytics: Fospha, Measured and INCRMNTAL are named as full-funnel and advanced measurement partners, while incrementality experiments are running with Haus, Measured and WorkMagic — the tooling advertisers typically demand before shifting budget from established search and social channels.
• Early results: According to DV Rockerbox, WeightWatchers recorded a cost per acquisition on ChatGPT Ads 15.3 percent below its blended paid-search benchmark. WorkMagic reported 67 percent of incremental purchases for wellness brand Dose came from net-new customers.
• New-customer claim: Triple Whale found 93 percent of Portland Leather visitors arriving via ChatGPT Ads were new. OpenAI did not publish spend levels, campaign duration or the number of advertisers behind these figures, and the results are attributed to the measurement partners rather than independently audited.
• Brand safety: A Negative Phrases tool is being made available to qualifying advertisers, and OpenAI is piloting brand suitability evaluation with DoubleVerify and Integral Ad Science. Mark Zagorski, CEO of DoubleVerify, said context in ChatGPT “is defined by the conversation itself rather than a page or video.”
• Industry framing: Lidiane Jones, CEO of Integral Ad Science, said marketers moving into AI-native environments need assurance that brand safety protections are durable, consistent and able to operate at scale. Businesses can sign up through ads.openai.com.

Background:

OpenAI has relied principally on consumer and enterprise subscriptions for revenue. The company frames advertising as part of a broader effort to help users discover products and make purchase decisions inside ChatGPT itself.

Between the lines:

The measurement build-out is the commercial tell. Conversion integrations, ten attribution partners and incrementality testing are the infrastructure performance advertisers require before moving budget, which suggests OpenAI is positioning ChatGPT Ads against paid search rather than as an experiment. The WeightWatchers figure, benchmarked explicitly against blended paid search, points at the same target.

What’s next

Watch the US test launch later this month, which advertisers join the initial group, whether the format extends beyond image generation, and results from the DoubleVerify and Integral Ad Science brand suitability pilots.