OpenAI switched on "creative enhancements" for ChatGPT Ads by default, so the system can rewrite your headlines, translate your copy, resize your image, and even change the landing page your clicks go to — without previewing every version.
2026-10-31 · by Moe Ameen
OpenAI has turned on "creative enhancements" for ChatGPT Ads by default, letting its ad system automatically change a running ad to better match the conversation it appears in. OpenAI emailed advertisers about the change on October 20, 2026, saying the setting applies to new and existing campaigns "starting Friday, October 30, 2026." In the days before that, OpenAI's own help page briefly showed different effective dates, before settling on that same October 30 date to match the email. Either way, as of this writing the setting is live for any campaign that didn't opt out first.
A single "creative enhancement" toggle covers four kinds of change. The system can rewrite or write new headlines and descriptions to connect your offer to what the user is asking; translate your ad text into another language; resize your image to fit the ad format; and change the destination URL to send the click to a more relevant page on your own site. OpenAI says it draws on the conversation context or the user's interests plus your existing ad and landing-page content, and that variations may be prepared in advance or generated in real time. Critically, these variations can run without each one being reviewed or approved by the advertiser, and Ads Manager does not preview every version — a shift from the earlier text-customization option, which generated copy you still had to approve before it ran.
The control is account-level by default, with campaign-level settings overriding it, and any opt-outs you already saved at either level were preserved when the default flipped on. You can change the toggle at any time in the Settings section of Ads Manager — but because a campaign setting takes priority, an account-level opt-out does not automatically switch off a campaign where the feature is already on, so you have to check each campaign. OpenAI's own guidance also cautions on measurement: it says changes in results over time do not, on their own, establish the effect of creative enhancements, and it describes no report that separates an enhanced version from your original. Confirm the current toggle state, dates, and reporting on OpenAI's advertising help pages, since this is new and moving.
This change moves a decision you used to make — what your ad says, shows, and links to — inside OpenAI's optimizer. You can leave it on and accept the trade, or opt out per campaign and keep control. Either path rewards the same thing: a deep bench of on-brand variations you approved yourself. Opt out, and you still want enough tested creative to improve the ad by hand. Leave it on, and the raw material the system remixes is still your copy, your image, and your pages — better inputs, better output. The problem both ways is volume: producing many genuinely on-brand variations is the bottleneck, which is exactly why "let the platform rewrite it" sounds appealing in the first place.
That is the gap [Kompozy](/) closes. It's a generation-and-publishing engine, not an ad tool: from one [Persona Brief](/glossary/persona-brief) it produces a batch of on-brand assets — Photo Posts and Quote Graphics, brand-exact Carousel Posts, and short-form video like Persona Shorts and Marketing Shorts — every one held to the same voice and banned-word list, so a dozen variations stay recognizably yours instead of drifting. Upload the strongest as ad creative and let your own CPC data pick the winner, no auto-rewrite required. The same assets publish as organic posts across eight social platforms plus blog and email, and that consistent footprint is what earns the free mention when someone asks ChatGPT for a recommendation — [generative-engine optimization](/glossary/generative-engine-optimization) that no ad setting can toggle off. For the paid-channel playbook, see our [ChatGPT Ads](/news/chatgpt-ads) coverage.
They are an automatic optimization setting for ChatGPT Ads that lets OpenAI's system change a running ad to better match the conversation it appears in. A single toggle covers four changes: rewriting or writing new headlines and descriptions, translating the ad text, resizing the image to fit the format, and changing the destination URL to a more relevant page on your site. OpenAI turned it on by default for new and existing campaigns at the end of October 2026.
Yes. OpenAI emailed advertisers on October 20, 2026 that the setting would apply to new and existing campaigns by default, effective Friday, October 30, 2026 — the same date its help page lists. Any opt-out you had already saved at the account or campaign level was preserved, but campaigns without one inherited the feature switched on.
Open the Settings section of Ads Manager and turn the "creative enhancement" toggle off. The setting exists at both the account level (the default for all campaigns) and the campaign level, and the campaign setting takes priority — so an account-level opt-out will not switch off a campaign where the feature is already on. Check each campaign individually, and you can change the setting at any time.
Yes. Opt the campaign out and only the copy, image, and landing page you supplied will run. Because you still need enough on-brand variations to test and improve by hand, a content engine like Kompozy generates those assets — image and video — from one Persona Brief, which also keeps every variation consistent if you leave enhancements on.