<\/span><\/h2>\nThree implications deserve your full attention:<\/p>\n
Your landing page is now source material for copy you will never preview.<\/strong> Every page on your site is a potential input to ad variations that run without your sign-off. Thin pages, outdated offers, placeholder text \u2014 all fair game for the enhancement engine.<\/p>\nYou are solely responsible for what the AI writes.<\/strong> Section 5.3 of OpenAI’s Advertising Terms makes the advertiser “solely responsible” for each campaign, including AI-generated outputs \u2014 and it explicitly covers “any optional tool or feature that is enabled by default and remains enabled”. If the AI rewrites your headline into something non-compliant, that is your problem, not OpenAI’s. For regulated verticals \u2014 legal, finance, healthcare \u2014 that sentence should decide your setting for you.<\/p>\nThere is a reporting blind spot.<\/strong> Ads Manager reports impressions, clicks, spend, CTR and attributed conversions at campaign, ad group and ad level \u2014 but there is no report showing each enhanced version individually. And OpenAI adds its own caveat: changes over time “do not, on their own, establish the effect of creative enhancements”. You can see performance move; you cannot cleanly attribute the move to the feature.<\/p>\nNone of this is hypothetical. If you run ChatGPT Ads and have not touched the setting, enhancements are on right now, rewriting against your landing pages as you read this.<\/p>\n
<\/span>The ChatGPT Ads Creative Enhancements Audit: What to Do This Week<\/span><\/h2>\n\n- Check your setting.<\/strong> In Ads Manager, open Manage settings and confirm whether creative enhancements are on at the account level \u2014 then check each campaign, because a campaign-level setting overrides the account default. An account-wide “off” does not guarantee every campaign is off.<\/li>\n
- Make the call per campaign.<\/strong> Turn it off for compliance-sensitive, brand-critical and high-spend campaigns. Consider leaving it on for exactly one low-risk campaign as a controlled test \u2014 the campaign-level override makes this possible without betting the account.<\/li>\n
- Audit your landing pages.<\/strong> Every page is now potential ad copy. Fix thin content, outdated pricing, expired offers and placeholder text before the AI starts quoting them to customers.<\/li>\n
- Protect your tracking.<\/strong> OpenAI’s ad creation guide tells advertisers to add UTM tags to each ad’s destination URL \u2014 but the help page does not explain how those parameters survive when the AI changes the destination. Verify in analytics that attributed traffic still carries your tags; if it does not, that alone is a reason to opt out.<\/li>\n
- Set a review cadence.<\/strong> Watch CTR and conversion rate at the ad level weekly. Remember there is no version-level report, so treat sustained shifts after October 30 as your signal \u2014 and keep your test campaign isolated so the signal stays readable.<\/li>\n<\/ol>\n
<\/span>Should You Leave It On?<\/span><\/h2>\nIt depends on what the AI has to work with. If you cannot audit every landing page before the AI starts reading them, a blanket opt-out is the safe default \u2014 off at the account level, then selectively test. If your pages are tight, your vertical is unregulated and your tracking is verified, a controlled test on one campaign may be worth running. OpenAI calls the feature “designed to” improve performance and publishes no performance data, so treat the effect as unknown until you measure it yourself.<\/p>\n
One reassurance from the help page, for the skeptics: creative enhancements do not change ChatGPT’s answers. Ads are separate from answers and clearly labeled \u2014 the automation touches your ad, not the model’s response.<\/p>\n
The broader lesson is the one Meta taught last month: on AI ad platforms, the default is the product decision. Whoever controls the default controls your creative \u2014 unless you go and turn it off.<\/p>\n
<\/span>Frequently Asked Questions<\/span><\/h2>\n<\/span>When did ChatGPT Ads creative enhancements switch on?<\/span><\/h3>\nOctober 30, 2026, for new and existing self-serve and BPO campaigns. OpenAI’s advertiser email gave Friday, October 30; the help page initially said Thursday, October 29 before settling on October 30. Sales-managed accounts are not affected.<\/p>\n
<\/span>How do I opt out of ChatGPT Ads creative enhancements?<\/span><\/h3>\nIn Ads Manager, select Manage settings and turn the setting off \u2014 at the account level, the campaign level, or both. Campaign-level settings take precedence over the account default, and existing opt-outs at either level are preserved. You can change the setting at any time.<\/p>\n
<\/span>Can I allow image resizing but block headline rewrites?<\/span><\/h3>\nNo. Creative enhancements are a single toggle covering all four changes \u2014 rewritten headlines and descriptions, image resizing, destination-URL optimization and translations. You cannot enable some and disable others.<\/p>\n
<\/span>Will I see each enhanced version before it runs?<\/span><\/h3>\nNo. OpenAI says enhanced ads “can run automatically without requiring you to review or approve each version”, and Ads Manager does not provide a preview of every variation. There is also no report that breaks out performance by individual enhanced version.<\/p>\n