Skip to content
Techtember

Always First. Fast Tech News.

Smartphone with ChatGPT on a background of U.S. dollar banknotes
Photo: Jernej Furman / Flickr via Wikimedia Commons (CC BY 2.0)

Opinion

Opinion: ChatGPT Is Putting Ads Where You Go to Imagine Things

OpenAI's new visual ads will appear while ChatGPT generates the image you asked for. The labels are fine. The placement is the real product.

Los Angeles · Marcus Reid, Opinion · Oct 6 2026

LOS ANGELES - OpenAI announced Monday that ChatGPT will start showing a new kind of ad, a visual one, and it picked a revealing place to put it. The first test will run during image generation, the moment when someone types out the kitchen they want, the outfit they are picturing or the trip they are planning, and waits for ChatGPT to draw it. Testing starts later this month in the United States with an initial group of advertisers, the company said.

OpenAI describes the format plainly. It uses “images showing product inspiration, product usage, or the experiences they make possible” to help people “imagine how products and services could fit into their lives.” The ads will be clearly labeled and kept separate from the image being created, and OpenAI repeated that advertising “does not influence the answers ChatGPT provides.” I take the company at its word on all three points. The trouble is that none of them is the point.

Hand holding a smartphone running ChatGPT in front of a laptop Photo: Jernej Furman / Flickr via Wikimedia Commons (CC BY 2.0)

Think about what a person is doing when they ask an AI to draw something. They are describing a want in their own words, often in more detail than they would ever type into a search box. A prompt for a cozy reading nook is a shopping list with feelings attached. An ad that appears right then, showing a real sofa in a real room, does not need to touch the generated picture to borrow its mood. It may be the most intimate ad slot in consumer tech, and OpenAI is pitching it to marketers as exactly that.

The rest of Monday’s announcement confirms who the audience is. OpenAI said ChatGPT reaches 1.2 billion people a week and called it “the world’s largest AI-native consumer platform.” It added integrations with Hightouch, Tealium and LiveRamp so advertisers can pipe in conversion data, a long list of attribution partners, and incrementality experiments with Haus, Measured and WorkMagic. It shared early wins too. According to DV Rockerbox, WeightWatchers paid 15.3 percent less per acquisition on ChatGPT Ads than its blended paid-search benchmark, OpenAI said.

Smartphone showing ChatGPT against a U.S. flag Photo: Jernej Furman / Flickr via Wikimedia Commons (CC BY 2.0)

None of this is a scandal. Free services need money, and OpenAI introduced ads earlier this year to support its free and low-cost tiers before expanding them to India in August, TechCrunch reported. TechCrunch also made the obvious point that ads inside the image tool could push casual users toward a paid plan to avoid them. In Europe, where ads reached 31 countries in late August for Free and Go users, Plus, Pro and Enterprise subscribers do not see them, according to The Next Web.

There are real guardrails, and they deserve credit. OpenAI says its placement checks assess whether a conversation is emotionally vulnerable or sensitive before any ad appears, and it is building brand suitability pilots with DoubleVerify and Integral Ad Science that test those rules without access to real conversations. But those protections are written for advertisers’ comfort as much as users’. The new Negative Phrases tool, for example, lets qualifying brands keep their ads away from topics they dislike. Monday’s announcement offered users no equivalent switch.

Hand holding a smartphone with the ChatGPT app open Photo: Jernej Furman / Flickr via Wikimedia Commons (CC BY 2.0)

Europe will at least get to see the receipts. The European Commission designated ChatGPT a very large online search engine on Aug. 31, The Next Web reported, and from January OpenAI must keep a public repository of a year of ads with the advertiser, run dates, targeting parameters and reach. American users, who get the image ads first, have no such window.

Here is a fair ask. If OpenAI is confident the visual ads help people, it should publish that same repository worldwide, and give every user, paying or not, a setting to keep ads out of image generation. The company says ChatGPT’s answers stay independent. The canvas should stay that way too.