ChatGPT Ads now rewrites your ads by default

What happened

OpenAI emailed ChatGPT advertisers on 20 October to say a feature called creative enhancements would be switched on by default. It went live around 30 October for new and existing campaigns that had not opted out. Trade outlets including Search Engine Land, Search Engine Journal and PPC Land reported the change.

With the feature on, ChatGPT can adapt an ad to the conversation it appears next to: it may reword the headline and description, translate the text, resize the image or send the click to a different page of the advertiser's website. OpenAI's help page says these versions can run without the advertiser reviewing each one, and there is no preview of every variation.

The key details

• What can change: headlines, descriptions, image size and crop, ad language, and the landing page on your own site.
• Start date: 30 October 2026 in OpenAI's email. OpenAI's own pages have also shown 29 and 31 October, so check your account.
• Who it applies to: campaigns in the self-serve ChatGPT Ads Manager that did not opt out. Reports differ on whether campaigns bought through agency partners are included.
• Approval: enhanced versions can run automatically, without approval of each version.
• How to opt out: in Ads Manager, open the settings and turn creative enhancements off for the account, or for a single campaign. A campaign setting overrides the account default, and earlier opt-outs are kept.
• Also new: multi-image ads, in beta since 26 October, are now open to all advertisers, according to OpenAI's advertiser email.

The background

Ads in ChatGPT began as a pilot in February 2026. In May, OpenAI opened a self-serve Ads Manager in beta with cost-per-click bidding, and it has since opened to more US businesses. Earlier this month it also started testing visual ads next to images people generate in ChatGPT (see our story on ChatGPT visual ads).

Automatic AI changes to ads are becoming the norm across platforms. Meta's updated terms, also starting 30 October, make advertisers responsible for whatever its AI changes in their ads (see our Meta story).

What it means for brands

If you run ChatGPT ads, open Ads Manager today and decide on purpose whether ChatGPT may rewrite your ads. For regulated categories, price claims, product names that must not be translated, or a strict brand voice, switching it off and testing your own versions is the safer choice.

The bigger lesson is the same on every platform: let AI help you make more versions, but make and approve them yourself before they go live. That is how we work on AI video ads: many finished variations, every one checked by the brand, instead of a platform changing your creative after launch.