For most of ChatGPT’s life, the wait while an image generated was dead time. You typed a prompt, watched a progress box fill, and got your picture. OpenAI has now decided that wait is worth money. Later this month it will start testing visual ads that appear right under that progress box, and it says ChatGPT now reaches 1.2 billion people every week.
The short version
- OpenAI will test visual ads during image generation in the US, starting later in October with an initial group of advertisers
- The ads sit below the progress box, are clearly labeled and are kept separate from the generated image
- The test targets the free tier, and paid plans such as Plus, Pro and Enterprise are reported to stay ad-free
- Brand safety and measurement come from DoubleVerify, Integral Ad Science, Kochava, AppsFlyer and Triple Whale
What the new ad format actually is
Until now, ChatGPT ads were text-led. This is the first format built around a moment that is already visual. When a free user in the US asks ChatGPT to create an image, a picture-based sponsored unit appears underneath the progress indicator while the image renders. Once the image is ready, the user gets their result as usual.
OpenAI is stressing three promises. The ad will be labeled as an ad, it will be kept apart from the generated output, and it will have no influence over what ChatGPT says or draws. That last point is the one that matters most for trust, and we will come back to it below.
| Detail | What has been reported |
|---|---|
| Where it appears | Under the progress box during image generation |
| Who sees it | US free-tier users in the test |
| Who does not | Plus, Pro and Enterprise subscribers |
| When it starts | Later in October, with an initial advertiser group |
| Labeling | Clearly marked as an ad and separate from the image |
| Brand safety partners | DoubleVerify and Integral Ad Science |
| Measurement partners | Kochava, AppsFlyer and Triple Whale |
Why image generation is the obvious place to start
Image generation is a good ad slot for a plain reason: the user is already waiting and already looking at the screen. A text answer streams in and holds attention. An image takes a few seconds to render and leaves a gap. A banner dropped into that gap does not interrupt anything the user was reading.
It also gives advertisers something text chats do not. A prompt like “a cozy living room with a green sofa” tells you a lot about what someone might want to buy. OpenAI says the ads will not change the output, but the surrounding context is clearly valuable to anyone selling furniture, travel or fashion.
The guardrails, and what they leave open
OpenAI says it is working with DoubleVerify and Integral Ad Science to keep ads out of emotionally vulnerable, sensitive or personal conversations. Placement checks look at whether a conversation is unsuitable before an ad appears. The safety partners are said not to read private conversations while verifying that the guardrails work.
That is a sensible starting list, but it leaves open questions that only the test will answer. How does a system judge that a prompt is sensitive when the prompt is a request for a picture? Will a person generating a memorial image for a lost pet see a sponsored banner? OpenAI has not published a list of excluded categories in the reports we reviewed, so treat the promise as a design goal, not a measured result.
Measurement is where the money is
The more interesting half of the announcement is about proof. Advertisers will be able to use Kochava, AppsFlyer and Triple Whale to see whether a click inside ChatGPT turned into an install or a purchase. That is the same attribution plumbing that powers mobile and ecommerce ad buying elsewhere, and it is the thing that makes a new ad surface credible to a marketing budget.
If you follow how OpenAI has been building this business, the pattern is steady. We covered the earlier push when OpenAI put sponsored agents inside ChatGPT chats, and the ad business has reportedly crossed a $1 billion annualized run rate roughly 200 days after launch. Ads also expanded to India in August. Visual ads are the next layer, not a new direction.
What this tells us
- Free is becoming ad-funded. With 1.2 billion weekly users, subscriptions alone do not cover a free tier that big
- Paying removes the ads. Keeping Plus, Pro and Enterprise clean gives OpenAI a simple upgrade pitch
- Attribution decides adoption. Advertisers follow measurable results, which is why the tracking partners matter as much as the format
- Trust is the real risk. The promise that ads never influence answers is easy to state and hard to audit from outside
The scale behind the move
The audience number is the reason any of this is possible. When ChatGPT passed a billion weekly users in August, we noted it was still losing share to rivals even as it grew. Two months later OpenAI is citing 1.2 billion. A company that big, with enormous compute bills, needs revenue that scales with usage rather than with the share of people willing to pay.
There is also a privacy angle worth watching. OpenAI has been open that real chats can be read by humans in some cases, as we explained when contractors were found reviewing real ChatGPT conversations. Advertising raises the stakes on how conversation data is used, even when the company says ads do not touch the answers.
What to do if you use ChatGPT
- If you are on the free tier in the US, expect to see a labeled banner during some image generations from late October
- If you pay for Plus or above, nothing changes according to current reports
- If you run a business, the early advertiser group is small, so watch for broader access and for how performance data looks after the first weeks
- If you care about privacy, review your data controls in ChatGPT settings before the test begins
The takeaway is simple. The loading bar is now ad space, the free tier is now an advertising product, and the proof that it works will come from the attribution tools, not from OpenAI’s promises.
Sources and further reading
- TechCrunch: OpenAI launches visual ads that appear alongside image generation results
- BleepingComputer: OpenAI will show visual ads in ChatGPT while you generate images
- Android Headlines: ChatGPT free tier is getting visual ads during AI image generation
- PYMNTS: OpenAI linking visual ads with ChatGPT images
About this article: GeekBlog covers U.S. technology news, AI, phones, smartwatches and gaming. Every story is written and checked under our Editorial Policy. Spotted a mistake or have a story tip? Contact our editors.
