OpenAI to test sponsored product images inside ChatGPT

OpenAI is expanding the advertising business inside ChatGPT from plain text placements to actual images. According to The Verge, the company will begin testing a visual ad format in the United States later this month, and the sponsored images will show up first in conversations where users ask ChatGPT to generate pictures.

From logos and links to product imagery

The move marks a notable escalation from the ad format OpenAI introduced in February. As The Verge explains, those earlier placements were modest: a company's name, logo and a link to its product, presented in a sponsored strip underneath the chat rather than inside the conversation.

The new format goes further by attaching images of sponsored products and services to the user's screen. According to the report, the first placements will surface when a user is generating images with ChatGPT. OpenAI says the sponsored image will stay separate from whatever the model actually creates, and the company states that ads will not affect the answers ChatGPT produces.

There is one significant carve-out: paying customers will not see any of this. The Verge reports that subscribers on ChatGPT's Plus, Pro and Enterprise plans are excluded from ads entirely, which keeps advertising limited to free users for now.

Guardrails around sensitive conversations

OpenAI acknowledges the obvious risk of inserting commercial content into a personal assistant. According to The Verge, the company says it has guardrails designed to judge when a conversation is appropriate for advertising, with the stated goal of keeping ads away from emotionally vulnerable, sensitive or otherwise unsuitable contexts.

That framing suggests OpenAI is aware that an ad served next to a grief-related query or a medical question could damage trust in the product. How reliably those guardrails work in practice will likely determine how the format is received.

The rollout is deliberately narrow. OpenAI is testing the visual ads with what it describes as an initial group of advertisers, and it reportedly intends to experiment with additional ad formats going forward.

Why it matters

ChatGPT has grown into one of the most heavily used consumer AI products largely without conventional advertising, so any expansion of ads changes the character of the service. The shift from a text-only sponsored strip to visible product imagery puts brand content much closer to the core experience, and adjacent to output the model itself generated, which raises fresh questions about whether users will clearly distinguish paid placements from organic results.

The subscriber exemption also formalizes a two-tier product: free users see ads, paying users do not. That gives OpenAI a second revenue stream beyond subscriptions at a time when running frontier AI models at ChatGPT's scale is expensive, and it signals that advertising is becoming a durable part of OpenAI's business rather than an experiment.

For advertisers, placement next to image generation is attractive real estate, since users in those sessions are often exploring products, designs or ideas. For users, the development means the answer surface of ChatGPT increasingly resembles a media platform with inventory to sell, and the quality of OpenAI's contextual guardrails will be tested in public as the US trial gets underway.