OpenAI launches visual ad format and measurement tools in ChatGPT

By Splitfeed EditorialLast updated Oct 5, 2026, 14:50 UTC

On October 5, 2026, OpenAI formally announced a new visual ad format inside ChatGPT, alongside expanded measurement tools, attribution partnerships, and brand suitability controls aimed at advertisers. According to Search Engine Roundtable, the visual ads are being tested specifically during AI image generation sessions. The same day, analytics reporting confirmed that Razorpay entered a partnership with OpenAI to extend ChatGPT advertising into the Indian market, signaling that the rollout is being treated as a scalable, international business from the start.

The move matters because it marks OpenAI's clearest step yet toward building a structured ad business inside a conversational AI product, a context that has no real precedent at this scale. Search advertising was shaped by the logic of queries and links, but ads inside an AI that generates images and holds extended dialogues raise genuinely open questions about where a placement begins and ends, how attribution works when a user does not click a traditional link, and how brand suitability can be enforced when outputs are generative by nature. The measurement and attribution tools OpenAI announced are partly an answer to those questions, but their adequacy is what advertisers and critics will likely test in the months ahead.

The Pro-AI camp has not yet published reactions to this story. That camp would typically argue that advertising revenue is what allows OpenAI to keep ChatGPT accessible and to fund further research, framing ads as a reasonable trade-off rather than a compromise. The Anti-AI camp is considerably less measured. BLKNSLVR described advertising as a parasite that has now latched onto AI as its newest host, framing the move as an extension of a broader commercial colonization of digital life. A separate account, tobellz.bsky.social, pointed to the quality problem embedded in the model, suggesting that AI-generated ad visuals can produce, in their words, grotesque nightmare fuel rather than the polished imagery advertisers expect, a concern that cuts at the credibility of the format itself. The Middle Ground camp, represented here by Can at canyesilyurt.com, takes a more resigned tone, accepting ads in ChatGPT as inevitable while acknowledging that conversations are about to feel noticeably more commercial.

The details worth tracking are how OpenAI defines and enforces brand suitability in a generative context, whether the Razorpay partnership produces a measurable advertiser base in India, and whether the image-generation ad tests expand beyond that session type into standard conversational responses.