Build1 publisher2 min readPublished
OpenAI writes itself discretion to remove ads that touch its competitive position
The stated grounds for declining, restricting or removing a ChatGPT advertiser include conflict with OpenAI's own business interests, and a dev.to summary reports the clause is already being applied to rival image and voice generation tools.
The Engineer · Build desk
What happened
- A dev.to post says industry reporting indicates the policy is being applied to campaigns for rival image-generation and voice-generation tools.
- The same post reports that video ads remain eligible while image- and audio-centric advertising faces restrictions.
- OpenAI announced initial testing of the advertising program in January 2026, with later expansion to partner ecosystems and measurement capabilities during the year.
Compiled by The EngineerSomething wrong?How this is made
Why it matters
- constraint Because removal sits alongside refusal in the policy, an accepted campaign is a revocable state, so a media plan cannot treat approval as a fixed input for the quarter.
- decision A team selling image or voice generation has to settle eligibility before it spends on creative. The deciding criterion is OpenAI's own commercial position, and it cannot be checked from outside.
- exposure Demand generation concentrated on the ChatGPT pilot is exposed to a review nobody else's policy governs, and the assets have to stay usable on channels that will still run them.
- contradiction The discretion clause is quoted policy text. The claim that image and audio tools are actually being restricted comes from reporting the post does not name, and that gap sets how confident any single category can be.
"Decline, restrict, or remove" covers three events with three different costs to an advertiser. A refusal at review costs you the creative you already produced. Removal costs you a campaign that was running, after the media commitment. According to a dev.to summary of OpenAI's advertising policies, the company reserves the right to do all three, and to refuse advertising content [1].
The stated grounds are the part that is hard to plan against. They include conflicts with OpenAI's "advertising principles, business interests, or competitive position" [2]. Principles are a published standard an advertiser can check its creative against. Business interests and competitive position belong to OpenAI and are assessed by OpenAI. The post says no complete public list of approved product types exists, so final eligibility depends on OpenAI's assessment [7].
The category-level claim rests on thinner evidence than the policy text. That post says industry reporting indicates the policy is being applied to rival image-generation and voice-generation tools [4], and that reports have video ads remaining eligible while image- and audio-centric advertising faces restrictions [6]. The post does not name the outlets carrying that reporting. It also says the available information does not describe a blanket ban on AI-related advertising [5].
For a team selling a generation tool, the check is narrow. The question is whether the product directly overlaps the image-generation or audio-generation capabilities OpenAI may treat as competitive [14]. If it does, eligibility has to be settled before any creative-production or media budget goes to ChatGPT placements [10]. The policy covers advertising inside ChatGPT and reaches no further: search, social and display run their own policies and review systems [12].
Ad eligibility is one question. Use of ChatGPT in research, content operations, customer support or internal workflows is another. The post says the ad policy does not change those uses [9]. It advises keeping paid acquisition planning apart from ChatGPT-based operational workflows [15].
One note on provenance. The account is a single dev.to post, and it closes by promoting Scalevise's AI Visibility and GEO Checker [13]. Its planning advice is checkable against the policy language it quotes: a paid-media plan resting on one new AI platform has more exposure to policy change than one able to shift across established channels and owned audiences [11].
OpenAI announced initial testing of the advertising program in January 2026, with later expansion to partner ecosystems and measurement capabilities during the year [8].
What to watch
- Whether OpenAI publishes category examples or an approved-product list alongside the next format rollout.
- Whether any advertiser reports a live ChatGPT campaign being removed after acceptance, not just declined at review.
- Whether the reported video eligibility survives the expansion to partner ecosystems and measurement.