Skip to content

SWLR Articles

Industry News: ChatGPT Ads turns creative rewriting on by default

Enterprise advertisers need an explicit governance decision because ChatGPT may change approved claims, translations, imagery and landing destinations without per-version review, affecting compliance, brand control and measurement.

Rob Kerry

TLDR;

OpenAI is making automated creative changes the default for new and existing self-serve ChatGPT Ads campaigns at the end of October. The system may rewrite copy, translate text, resize images and send clicks to a different page on the advertiser's site without approval of every version. Brands should decide whether that loss of control fits their legal, brand and measurement rules before the switch takes effect.

What happened

OpenAI told advertisers on 20 October that creative enhancements would be enabled automatically for new and existing self-serve campaigns. The email said the switch would begin on 30 October, while keeping existing account and campaign opt-outs. Advertisers can change the setting later, and a campaign setting takes priority over the account default.

OpenAI's product documentation says the system uses the conversation or user interests, the current advertisement and the landing page as inputs. It may create or update headlines and descriptions, translate text, resize images and choose a more relevant destination URL on the advertiser's site. Variations can be prepared in advance or generated in real time and can run without a preview or approval for each version.

The setting affects sponsored placements rather than ChatGPT's answers. OpenAI says ads remain separate and labelled, but that boundary does not remove the advertiser's responsibility for the message and destination a customer receives. The company has not described reporting that identifies each generated variation or isolates its performance.

PPC Land compared the conflicting activation dates in OpenAI's own materials. The advertiser email said 30 October, an earlier version of the help page said 29 October, and the page reviewed on 24 October now says 31 October. OpenAI has not explained the changes. Advertisers should therefore review the setting now rather than rely on one deadline.

Why it matters

AdExchanger highlighted the loss of per-version approval because this goes beyond simple image resizing. A rewritten headline can alter a regulated qualification, a translation can change the meaning of a promise, and destination optimization can route a click away from the page named in a media plan or legal review.

The destination change is also a measurement issue. Conversion tags, consent handling, product availability and campaign parameters need to work on every page the system may select. If an enhanced ad points to a category page instead of a product page, a higher click rate may still produce a worse customer journey or weaker conversion quality.

OpenAI frames the feature as a way to improve relevance and reduce creative work, but it has published no performance evidence for the automatic default. Account-level reporting can show whether results changed, yet it cannot prove the enhancement caused the change when bidding, audience mix, seasonality and destination pages also move.

For enterprise teams, the main change is governance. Paid media can no longer assume the approved asset is the only version customers will see. Legal, brand, localization, analytics and website owners need one decision about which campaigns may adapt, which claims must remain fixed and who is responsible for reviewing the results.

How your brand can benefit / be affected

Audit the account and campaign settings before the earliest date OpenAI has published. Opt out campaigns with regulated copy, contractual wording, fixed promotional terms or tightly controlled language. Keep the feature off when a destination URL must stay tied to a specific disclosure, price, stock position or measurement design.

For campaigns that remain enabled, clean the source material first. Remove expired offers, make prices and availability explicit, keep product claims consistent across the ad and landing page, and check that translations have approved terminology. The model can only adapt what it reads, so conflicting pages create more room for a plausible but incorrect message.

Expand tracking and quality checks beyond the original landing page. Validate consent, analytics and conversion events across eligible destinations. Monitor search terms, click quality, lead quality, conversion value, returns and complaints. Preserve a controlled group without enhancements if the account structure allows a fair comparison.

Search Engine Land's practical warning is to make the toggle a deliberate choice. Record the account default, campaign overrides, decision owner and review date. Ask OpenAI for variation-level reporting and a stable activation date. Until those controls exist, treat creative enhancements as an automated publishing system with commercial consequences, not a harmless formatting option.

News date: 20 October 2026. Editorial review: 24 October 2026.