OpenAI told advertisers on October 20, 2026, that it will switch on "creative enhancements" for their new and existing self-serve ChatGPT campaigns at the end of the month, allowing the system to rewrite headlines and descriptions, resize images, change destination URLs and translate ad text, and to run those versions without the advertiser approving each one. The notice, sent by email and first reported by Barry Schwartz at Search Engine Roundtable, gives Friday, October 30, 2026, as the start date. OpenAI's linked help page, updated within the past day, now says Thursday, October 29.

In Short

OpenAI is about to let its own system rewrite the ads that businesses run inside ChatGPT, changing the words, picture size, language and even which page a click lands on, so the ad fits the conversation it appears in. This affects every advertiser that buys ChatGPT ads through OpenAI's self-serve tool, because the setting is on unless they turn it off. After the switch, the version of an ad that a ChatGPT user sees may be one the advertiser never read, and advertisers who want to approve every word need to opt out in Ads Manager before the end of October.

What the email says

The source for the change is an email OpenAI sent to advertisers, including Schwartz, whose RustyBrick account received it. Search Engine Roundtable published the email as a screenshot alongside a transcription on October 20 at 3:35 pm, under the headline "OpenAI Auto Enabling Creative Enhancements In ChatGPT Ads."

The email opens with the date and the scope. "Starting Friday, October 30, 2026, we'll enable creative enhancements for your new and existing self-serve campaigns," according to OpenAI.

It then describes the purpose in a single sentence. "Creative enhancements make your ads more relevant to the conversation and/or user interests, helping improve performance, while reducing the work of creating multiple versions," the email says.

The third paragraph sets out the inputs and the four kinds of change. "We use your existing ad and landing-page content to adapt the way your message is expressed. Enhancements can include new or updated headlines and descriptions, image resizing, destination-URL optimization, and translation into other languages," according to OpenAI.

The fourth paragraph is the one advertisers have reacted to most. "When enabled, these ads can run automatically without requiring you to review or approve each version," the email states.

The final paragraph covers the opt-out. "You can opt out of creative enhancements before October 30, 2026 or change your settings at any time in Ads Manager. We'll preserve your existing account and campaign-level opt-outs, and campaign-specific settings will take priority over account defaults," according to OpenAI. The email closes with two links, "Manage settings" and "Learn more."

The four things that can change

Search Engine Roundtable's post reproduces OpenAI's breakdown of what the system may alter, which corresponds to the help document it links. There are four categories.

  • Headlines and descriptions. "OpenAI may create new or updated text to connect your offering to the conversation or user interests."
  • Language. "OpenAI may translate your ad text into another language."
  • Images. "Images may be resized to fit the ad format."
  • Destination URLs. "The destination URL may be optimized to direct people to a relevant page on your website."

According to the same passage, the system draws on three inputs: "the context of the conversation or user interests, your existing ad and content from the ad's landing page." Timing varies. "Variations may be prepared in advance or created in real time based on the conversation," the help text says, as reproduced by Search Engine Roundtable.

Two of the four categories go beyond rewording. Translation means an ad written in one language can be served in another, a capability that matters more now that ChatGPT ads are sold in 89 countries. Destination-URL optimization means a click may land on a page other than the one the advertiser entered, chosen by OpenAI from the advertiser's own site. Neither the email nor the help text quoted in the post explains how a replacement page is selected, whether it must sit on the same domain or path, or whether advertisers can exclude pages.

Image resizing is the narrowest of the four, and the text does not say whether resizing involves cropping. That distinction matters to brands with logos or text near the edges of an image.

Where the setting sits

The control is in the Settings section of Ads Manager, which still carries a "Beta" label in OpenAI's interface. A screenshot in Search Engine Roundtable's post shows a single card titled "Creative enhancement" with a toggle set to "On." The card's description reads: "Leverages your existing ad and landing-page content to adapt your message to the current conversation or user interests. This can include new or updated headlines, descriptions, resizing images, and optimizing destination URLs, as well as translations into other languages. Subject to the Ad Tool Terms."

Schwartz wrote that the setting was already toggled on in his account when he captured it. The screenshot links to the settings page at ads.openai.com.

The email describes two levels of control. An account-level default applies across campaigns, and campaign-level settings override it. Existing opt-outs at either level carry over. That structure means an advertiser who switches the account default off does not need to touch each campaign, but one who leaves the default on can still exclude individual campaigns. The email does not mention an ad-level or ad group-level control, and it does not allow advertisers to keep some of the four enhancement types while switching off others. The toggle in the screenshot covers all four together.

Two dates for the same switch

The email gives Friday, October 30, 2026. OpenAI's help article, "Creative enhancements for ChatGPT Ads," which both the email's "Learn more" link and Search Engine Roundtable's post point to, displayed an "Updated: 12 hours ago" line when PPC Land checked it on October 21. In that version, the automatic switch is dated Thursday, October 29, 2026, one day earlier than the email.

The help page also describes the scope differently. The email refers to "self-serve campaigns." The help page, as checked on October 21, refers to new and existing campaigns without that qualifier. It does not say whether campaigns bought through OpenAI's sales team or through technology partners are included.

OpenAI has not explained either difference publicly. For advertisers relying on the email's deadline, the practical effect is that the help page now puts the switch a day sooner.

What the help page adds

The help page, as checked on October 21, contains a few details that the email does not. It says the experience does not provide a preview of every variation. It directs advertisers to Ads Manager reporting at the campaign, ad group and ad levels, listing impressions, clicks, spend, click-through rate and attributed conversions, and it cautions that changes in results over time do not, on their own, establish the effect of creative enhancements. It asks advertisers to keep ads and landing pages current and to comply with OpenAI's Ads Policies. In a short FAQ, it restates that ads do not influence ChatGPT's answers and that ads are separate and clearly labeled.

The page does not describe any reporting that separates enhanced versions from the original creative, or a way to see which headline, language or landing page was shown in a given impression. That gap matters because the help page's own caution about reading changes over time implies that advertisers have no direct comparison to work from.

What neither source says

Set against the questions an advertiser would ask before leaving the setting on, the email and help page leave a number of gaps:

  • which model or system generates the new text and translations
  • whether "user interests" draws on ChatGPT memory, past conversations or other signals, and how that interacts with OpenAI's promise to keep conversations private from advertisers
  • whether enhanced versions pass through the same policy review as the original ad, or are checked only at the template level
  • whether any label tells a ChatGPT user that the ad text was generated or translated by OpenAI rather than written by the advertiser
  • which languages translation covers, and whether it applies in all 89 markets
  • how the replacement destination URL is chosen and whether pages can be excluded
  • whether the change applies to image-led formats in testing or to Sponsored Agents
  • any performance data from testing, such as click-through or conversion differences
  • whether the switch rolls out to all accounts on one day or gradually

The question of responsibility is also unaddressed. The settings card says the feature is "Subject to the Ad Tool Terms," but neither source summarises how those terms allocate liability if an automatically written headline or translation makes a claim the advertiser would not have made, in a regulated category or a language the advertiser does not read.

How it fits the ChatGPT ads build-out

Creative enhancements arrive eight months after ads began. When OpenAI said in January it would test ads for logged-in adult Free and Go users in the US, the format was a single sponsored unit placed below an answer and separated from it. The US pilot started on February 9.

The self-serve tool that the new setting lives in is newer. When OpenAI opened ChatGPT Ads Manager to all US businesses on May 5, each ad consisted of an advertiser name, a favicon, a title, description copy, a landing page URL and an image. OpenAI's documentation at the time recommended a headline of about 16 characters and a description of about 32. Campaigns used "context hints" at the ad group level instead of exact-match keywords, and OpenAI said the matching system also weighed landing page content, the ad title and the ad copy. The May documentation described account verification but no review step for individual ads.

Creative enhancements extend that logic. Under contextual advertising as OpenAI practised it from launch, the system chose which fixed ad to show based on the conversation. With enhancements on, the conversation can also shape what the ad says.

The scale at which this will happen has grown quickly. On October 21, OpenAI said ChatGPT ads are now sold in 89 countries and that more than 100,000 advertisers have run campaigns since launch, a cumulative figure without an active count. Reporting has not kept pace with targeting in other areas. In August, PPC Land found that ChatGPT Ads gained iOS, Android and web targeting while reporting stayed at two device buckets, with OpenAI saying only that more detailed platform reporting was planned.

Third parties have built similar systems on top of OpenAI's inventory. Criteo, a technology partner for ChatGPT ads, described its Prompt Smart Ads in June as automatically generating creatives and product descriptions tailored to the context of a user's query, drawing on catalog data and the conversational prompt. Criteo reported roughly four times higher spend after activation, a measure of budget delivery rather than results. OpenAI's version moves the same idea into its own self-serve tool and makes it the default.

The industry pattern of default-on automation

The approach follows a pattern set by the largest ad platforms, where generative and automated creative features have repeatedly been switched on for existing campaigns with an opt-out window.

Meta's Advantage+ creative enhancements are on by default. PPC Land reported in October 2025 that Advantage+ can generate up to 150 creative variations and that advertisers generally cannot preview every one, which allowed mismatched pairings of images and text to reach users. In August 2025, Meta made dynamic media the default for new Advantage+ catalog ads created through its API, with enforcement reaching 100% on October 20, 2025. Last week, PPC Land reported that Meta will turn on product browsing in existing ads on October 28 unless advertisers opt out, citing a 3% average conversion gain from an August experiment across about 117,700 advertisers. Media buyers objected to the timing ahead of Black Friday.

Google has done the same with video and search. In March, Google began adding AI-generated voice-over to existing Performance Max videos, with an opt-out deadline of March 20. The voice-over was saved as a new asset and the original video was not overwritten. In August, Google told advertisers that campaigns using automatically created assets would move to AI Max on September 1, with text customization switched on for that group. Final URL expansion, Google's closest equivalent to OpenAI's destination-URL optimization, was not switched on by that upgrade.

OpenAI's change bundles all of these levers into one switch. Text generation, translation, image resizing and landing page selection are controlled by a single toggle, set to on, with no per-type control described in either source. Google kept final URL expansion out of its automatic AI Max upgrade, and Meta's product browsing notice applied to one element of the post-click experience. The OpenAI setting is broader than either, and it arrives on a platform that is still labelled beta. Whether the system amounts to dynamic creative optimization in the conventional sense is unclear, since that usually assembles variants from components the advertiser supplied, while OpenAI says it may "create new or updated text."

Early reaction

Search Engine Roundtable collected reactions from paid search practitioners on X. Anthony Higman posted that he would switch ChatGPT ads off, describing the feature as "Pre-Opted In" and saying OpenAI was following another platform's playbook. In a second post, he wrote: "If An Ad Platform Is Framing Something As "Less Work For You" You Can Pretty Much GUARANTEE What They Really Mean Is : We Are Scamming You By Serving More Of Your Ads On Things You Dont Want To Show Up For." That is his opinion; OpenAI's email says the purpose is relevance and reduced workload.

Melissa Mackey, replying on October 20, cited her own results. "Yikes. We ended up pausing our ChatGPT ads test due to poor lead quality and now I'm glad we did. Not cool," she wrote.

The reactions are from two practitioners and do not represent a survey of advertisers. They do, however, focus on the same two points the email makes explicit: that the setting is on by default, and that versions can run without approval.

Why this matters for the marketing community

For advertisers in regulated categories, such as financial services, the absence of per-version approval is the central issue. A headline rewritten to fit a conversation, or translated into a language the compliance team does not read, carries the same legal exposure as one the advertiser wrote. Neither the email nor the help page describes a safeguard specific to those categories.

For performance marketers, the issue is measurement. ChatGPT ads already report at a coarser level than they target. Adding automatically generated headlines, translations and landing pages, without reporting that shows which version ran, makes it harder to attribute a change in click-through or conversion rate to anything specific. OpenAI's own help page concedes the point by warning that changes over time do not establish the effect of enhancements.

For brands, the change alters what an ad in ChatGPT is. Until now, an advertiser could know the exact words a user saw. After October 29 or 30, depending on which OpenAI source is correct, that will no longer be true for any account that leaves the default in place. When Meta set its own October 28 default, media buyers objected as much to the timing and the short notice as to the feature itself. OpenAI's email gave advertisers ten days; its help page now gives nine.

Timeline

Summary

Who: OpenAI and advertisers running self-serve ChatGPT campaigns. The change was reported by Barry Schwartz at Search Engine Roundtable, with reactions from practitioners Anthony Higman and Melissa Mackey.

What: OpenAI will enable creative enhancements by default for new and existing campaigns. The system can rewrite headlines and descriptions, translate ad text, resize images and change destination URLs, using the advertiser's ad and landing page content plus conversation context or user interests, and the versions can run without per-version approval. Advertisers can opt out at account or campaign level in Ads Manager, and existing opt-outs are kept.

When: The email was sent and reported on October 20, 2026. The email gives Friday, October 30, 2026, as the start date; OpenAI's help page, updated on October 21, gives Thursday, October 29.

Where: In ChatGPT ads, controlled from the Settings section of OpenAI's Ads Manager beta. Neither source limits the change to specific markets.

Why: OpenAI says the feature makes ads more relevant to the conversation and user interests and reduces the work of creating multiple versions. It has published no performance data, reporting by version or label wording for enhanced ads.