OpenAI Posts the Roles That Will Build ChatGPT’s Next Ad Formats

OpenAI has posted three engineering roles to build ChatGPT’s next wave of ad formats, covering text, image, video, native, conversational, and interactive experiences. Each role pays $230,000 to $385,000 plus equity and is based exclusively in San Francisco. All three are hiring to build formats that don’t yet exist, with privacy and safety framed as engineering requirements baked into every rendering rule.

Three Listings, One Foundational Role

The three postings appeared this week on the Ad Formats career listing, as first reported by Digiday. The most senior is a Software Engineer, Ad Formats role requiring at least seven years of experience and labelled a “foundational role” across the full technical stack.

Two companion roles focus on iOS and Android respectively, each asking for at least four years. All three sit inside OpenAI’s newly formed Monetization team, described in the senior listing as a “new cross-functional group” building “user-first, privacy-preserving monetization products, including next-generation ads experiences.” Compensation for every role is $230,000 to $385,000 plus equity, with relocation assistance offered to bring new hires to the company’s San Francisco headquarters.

What the Engineers Will Build First

The senior role’s job spec lists the formats the engineers will build: text, image, video, native, conversational, and interactive. The senior engineer will own what the posting calls the “creative rendering and presentation layer of OpenAI’s ads ecosystem,” spanning every OpenAI surface that may eventually carry an ad. The iOS and Android engineers will build “policy-aware UX patterns” and “format validation” tooling specific to mobile.

The three roles share a mandate to ship systems that “render reliably, perform efficiently, and feel natural within the core ChatGPT experience,” per the listing. Rob Webster, CEO of TAU Marketing Solutions, framed the deeper work in comments on the new roles: tackling attribution, brand safety, and device modelling.

The senior listing names ad infrastructure that touches “ChatGPT and other OpenAI product experiences.” The engineer hired for the role is expected to “propose new directions, and help shape the roadmap for OpenAI’s ads formats platform.” That this work is being scoped at all means the company’s ad formats will extend well beyond the single unit now live.

Where ChatGPT Ads Live Today

ChatGPT advertising sits only on the free and Go tiers today. Pro, Business, and Enterprise subscriptions do not include ads, per OpenAI’s policy. ChatGPT Go launched in 171 countries since August and is priced at $8 USD/month. Ads run “at the bottom of answers in ChatGPT,” per the company’s ad principles, only when there is a relevant sponsored product or service tied to the conversation. OpenAI has also said users may “be able to directly ask the questions you need to make a purchase decision” from inside an ad, a hint that conversational advertising is on the product roadmap.

Ad tests have so far been limited to logged-in adults in the United States. An OpenAI update on May 7, 2026 said the company plans to “expand the ads pilot in ChatGPT in the United Kingdom, Mexico, Brazil, Japan.” That makes this week’s new engineering roles a literal build for what is about to go live in more places.

• United States (pilot underway)
• United Kingdom (planned expansion)
• Mexico (planned expansion)
• Brazil (planned expansion)
• Japan (planned expansion)

The Standard Unit Barely Moved

For all the hiring, the format OpenAI ships today is minimal: a headline, a short description, an image, and a link. The team has spent four months iterating on the balance of text and image so as not to crowd the conversation.

The most recent change is small. The unit’s width moved from 480 pixels to 440 pixels. Last month, Digiday viewed mockups of a wider iteration with a larger image and an optional personalised call-to-action button, but that mockup has not yet shipped.

OpenAI VP of monetization Benji Shomair framed the slower pace at a press roundtable in May, calling “creative variation” a “real key to success.” OpenAI ads boss David Dugan told Digiday last week that the team has been heads-down on getting the basics right, with the roadmap shaped by feedback from test advertisers helping to define what comes next.

• 480 to 440 pixels: the unit’s current width, narrowed from the original format
• 4 months: how long the team has been iterating on the live standard unit
• 6 new format types: named in this week’s Ad Formats job spec

Why Privacy Sits in the Job Spec

All three Ad Formats listings require engineers to “uphold the highest levels of safety, privacy, fairness, and policy compliance across all ad rendering and delivery systems.” The iOS and Android roles push the privacy mandate deeper, listing the build of “policy-aware UX patterns” and “format validation” as core duties. Privacy here lives inside the rendering rules.

OpenAI’s own principles guiding ChatGPT advertising align with that posture. Ads are “always separate and clearly labeled” from organic answers, the company has written. Conversations stay private from advertisers, and user data is never sold. Users can clear the data used for ads or turn off personalization.

The tension inside that stack is internal. Two of the commitments pull against the third: relevance to a conversation requires reading it. Privacy requires not doing so.

Are you optimizing for user trust or advertiser value? Those objectives are often incompatible. Solving that problem, if it’s solvable at all, will take more than innovation around ad formats.

Andrew Frank, research vice president and analyst at Gartner, named that tension in comments to Digiday as the “dual-alignment problem.” He cast doubt on whether engineering ambition alone can resolve it.

Two Speeds in the Ad Build

The Ad Formats hiring sits inside a ChatGPT ad business running on two tracks. The first is the slow standard unit, where each visual tweak is tested and weighed before it ships. Shomair’s “creative variation” framing in May was the public read on that pace, and the unit’s narrow recent trim fits the pattern.

The second is the engineering build behind this week’s postings, where OpenAI is hiring to define how every future format will render and what rules they will have to obey. The two tracks can coexist because they target different problems. What users see today is the narrow standard unit; what comes next is being defined by the engineers being hired to build it. Asked last week what comes next for ad formats in ChatGPT, OpenAI ads boss David Dugan was vague, and OpenAI did not respond to Digiday’s request for comment on the new roles.

Where the Stakes Sit

How big those stakes are is not lost on the market. Barclays projects ChatGPT ad revenue at $102 billion in ChatGPT ad revenue by 2030, against a backdrop where AI search advertising is forecast to cross $100 billion by 2030.

That scale is what the three engineering roles are being hired to support. The live format itself has barely moved, which is part of why the new roles exist.

So far [OpenAI] has just been plowing ahead with a global launch of the one single format and placement they’ve ever tested, without any idea if that actually works best for advertisers or for users.

Nate Elliott, eMarketer’s principal analyst for AI in marketing and commerce, shared that view with Digiday. The Ad Formats roadmap is shaped by what test advertisers ask for next, per ads boss David Dugan. Whether the rendering layer can carry both the new ad load and the privacy guardrails without one eroding the other will decide if the standard unit’s trim ends up looking like a footnote.