OpenAI’s ChatGPT Ads platform ships a feature called Text Customization, and it is turned on by default for advertisers who see it. The toggle lets AI rewrite and translate your ad headlines and descriptions automatically, and the on-screen warning says those versions “may be shown without your individual review or approval.” That detail, reported independently by paid-media analyst Jon Loomer, sits awkwardly next to OpenAI’s own announcement, which describes the same feature as something advertisers “opt into.”
This matters right now because Search Engine Roundtable flagged the gap on October 20, 2026, and advertisers running early ChatGPT Ads campaigns are the ones finding out the hard way. Here is what ChatGPT Ads Text Customization actually does, why the default setting contradicts OpenAI’s own framing, and how to turn it off if you want full control over what your ad copy says.
What ChatGPT Ads Text Customization Actually Does
OpenAI introduced the feature in a company blog post titled “Reimagining advertising with AI,” published September 16, 2026. The post describes an option that, “if enabled, adapts an advertiser’s existing headlines and descriptions to better fit the context of a conversation.” It also automatically translates ad copy into a user’s preferred language.
In practice, that means ChatGPT can rewrite the words you submitted for a campaign before showing them to a specific user. A headline you wrote for a general audience might get reworded to match what someone was just discussing with the chatbot. A description written in English might appear in Spanish or Japanese without you touching the campaign.
Loomer’s report, published on his Quick Video Tip blog, includes a screenshot of the toggle as it appears inside the ChatGPT Ads Manager campaign builder. The exact wording reads: “AI automatically generates personalized and translated versions of your ad’s headlines and descriptions, which may be shown without your individual review or approval.”
That is a stronger claim than OpenAI’s blog post makes. The company’s own help documentation for creating ads and creative variations, at the time of writing, does not mention Text Customization at all. The only AI-assisted feature described there is a suggested ad draft tool, which explicitly “does not generate new copy or imagery with AI” and requires the advertiser to review and assign each draft manually.
Why ChatGPT Ads Text Customization Ships On By Default
OpenAI’s blog post uses opt-in language. It says the company is “launching the ability to opt into AI-powered text customization,” which reads as an advertiser-initiated action.
Loomer’s account of the live product contradicts that framing. According to his report, the toggle is on by default when advertisers create a campaign, and it has to be switched off separately for every single campaign. There is no account-level setting to disable it once and move on.
My take: this is a familiar pattern for anyone who has managed paid social campaigns over the last few years. Ad platforms tend to launch automation features with careful opt-in language in press materials, then ship the actual toggle pre-selected inside the product. The press release reads one way. The settings panel reads another.
There is also no preview available before the ad runs. Loomer calls the feature “mysterious” because OpenAI gives advertisers no way to see what a customized headline or translated description will actually say before it reaches a user. You are trusting the model’s output sight unseen, for every conversation it decides is a good match.
How ChatGPT Ads Compares to Meta’s Advantage+ Creative
If this setup sounds familiar, it should. Meta built a nearly identical feature years ago inside Advantage+ Creative, which can swap in different headlines, crop images, and add overlays automatically. Loomer notes directly that “similar enhancements are the target of the most common complaints from Meta advertisers,” and that history is worth knowing before you touch the ChatGPT Ads toggle.
The complaint pattern on Meta’s side has been consistent. Advertisers report that automated creative changes sometimes strip out brand-approved messaging, misrepresent a promotion’s terms, or apply a tone that does not match the brand voice. Legal and compliance teams in regulated industries have had to push back hardest, because an auto-generated line about pricing or availability can create real liability if it is wrong.
Google has moved in a similar direction with AI Max for Search campaigns, which also expands and adapts ad text using AI signals. The difference worth noting is that Google’s rollout has included more visible controls and preview options as the feature matured, something ChatGPT Ads has not caught up to yet given how new the whole platform is.
What this means for you: if you already have a policy for handling Meta’s Advantage+ Creative, the same reasoning applies here. Decide in advance which campaigns can tolerate AI-rewritten copy and which cannot, rather than discovering the answer after a bad headline runs.
Where ChatGPT Ads Text Customization Fits in OpenAI’s Bigger Push
The feature did not launch in isolation. OpenAI’s September 16 announcement bundled it with several other additions to ChatGPT Ads, and the context helps explain why a default setting might have slipped through without much scrutiny.
The company is testing Sponsored Agents, a format that lets a person start a conversation with a business-sponsored agent after clicking an ad. OpenAI says this is being tested with select advertisers in the United States only, so it has a much smaller footprint than Text Customization right now.
OpenAI also introduced an Ads Manager plugin that lets advertisers use natural-language prompts inside ChatGPT itself to create, update, and analyze campaigns. The same announcement covered a new HubSpot integration, letting businesses create ads and follow up on leads directly inside HubSpot using their existing CRM context, and a Shopify integration that lets US merchants run ChatGPT Ads campaigns using products already synced through Shopify Catalog.
Each of those additions is a meaningful step toward a more complete ads platform. Marketers already using HubSpot’s Breeze Assistant for other AI-assisted marketing tasks will likely find the CRM connection convenient. But the pace of these releases is also a reason to check every new toggle carefully. A platform shipping this many features at once is more likely to leave a default setting unexamined, and Text Customization appears to be exactly that kind of gap.
The Real Risk: No Preview, No Approval Step
The core risk is not that AI rewrites your copy. Plenty of marketers use AI to draft and refine ad copy every day. The risk is that ChatGPT Ads Text Customization can publish those rewrites without a human checking them first.
Three scenarios stand out:
- A regulated business (finance, healthcare, legal services) runs copy that must use specific, approved language, and an automated rewrite drifts from that language without anyone noticing.
- A limited-time promotion gets translated or reworded in a way that misstates the offer, the deadline, or the discount amount.
- A brand with a very specific tone, built over years of consistent copywriting, suddenly has AI-generated variants running that do not sound like the brand at all.
OpenAI’s own Ad Tools Terms place responsibility for creative accuracy on the advertiser, not the platform. The terms describe the AI-powered tools as generating, modifying, or translating ad materials, and they make clear that advertisers are responsible for reviewing output for accuracy and compliance before it runs. When there is no preview step, that review responsibility becomes difficult to fulfill in practice.
This is separate from OpenAI’s other AI creative feature, the “generate ads for you” draft option under “Add new ad” in Ads Manager. That tool pulls a title, description, and image from your website to prefill a new ad, but it stops at a draft. The interface tells you to “review, edit as needed, and activate when you’re ready.” Text Customization skips that step for the headlines and descriptions it rewrites on the fly.
How to Turn Off ChatGPT Ads Text Customization
Based on Loomer’s reporting and OpenAI’s own product language, here is the practical process for reviewing or disabling the setting:
- Open ChatGPT Ads Manager at ads.openai.com and go into an existing or new campaign.
- Look for the Text Customization toggle in the campaign creation or editing screen. Not every advertiser account sees it yet, since OpenAI appears to be rolling it out gradually.
- Check whether the toggle is already switched on. Based on current reports, assume it is unless you have confirmed otherwise.
- Turn it off for any campaign where brand language, pricing, legal claims, or promotional terms need to stay exactly as written.
- Repeat the check for every new campaign you launch, since the setting applies per campaign rather than account-wide.
- Document the decision internally so anyone else on your team managing the account knows the setting was checked and set deliberately, not left on by accident.
If your account does not show the toggle at all, that likely means OpenAI has not enabled the feature for your account tier or region yet. Check back as the platform expands, since OpenAI has been rolling out ChatGPT Ads features to more advertisers through the rest of 2026.
Who Should Use ChatGPT Ads Text Customization, and Who Shouldn’t
This feature is not universally bad. It solves a real problem for a specific kind of advertiser, and it is worth being honest about both sides.
It is a reasonable fit for: ecommerce brands running broad, non-regulated campaigns where a slightly reworded headline carries low risk. Advertisers who sell into multiple languages and do not have the budget for professional per-market copywriting may also get real value from the automatic translation, as long as they spot-check output periodically.
It is a poor fit for: any advertiser in a regulated category like finance, healthcare, legal, or insurance, where exact wording often has compliance implications. It is also a poor fit for brands with a distinct, carefully maintained voice, and for any campaign promoting a specific, time-sensitive offer where a misworded rewrite could create a customer complaint or a refund dispute.
In my experience advising on paid media accounts, the safest default is to treat any auto-enabled AI creative feature as off until proven otherwise for your specific use case. Turn it on deliberately for the campaigns where the upside is clear, rather than leaving a platform default in place without checking what it actually does.
Why This Matters
ChatGPT Ads is still a young platform, and OpenAI is clearly building out automation features quickly to compete with Meta, Google, and LinkedIn. That speed is reasonable from a product standpoint, but it creates a real gap between what a company blog post says and what the live product actually does by default.
What to do next: audit every active ChatGPT Ads campaign this week and confirm the Text Customization setting on each one. Do not assume your account matches OpenAI’s opt-in description until you have checked the toggle yourself.
- Check the toggle on every live campaign, including ones you launched before this report.
- Write down your brand’s rule for which campaign types can tolerate AI-rewritten copy.
- Compare notes with your Meta Ads and Google Ads AI Max settings, since the same review logic applies across all three platforms.
FAQ: ChatGPT Ads Text Customization
Is ChatGPT Ads Text Customization a paid add-on?
No. It is a setting inside the standard ChatGPT Ads Manager campaign builder, not a separate paid feature. It is part of using the Ads Manager tool itself.
Can I preview the rewritten ad copy before it runs?
Not currently, based on available reporting. The toggle’s own description states that personalized and translated versions “may be shown without your individual review or approval.”
Does Text Customization change images as well as text?
No. Based on OpenAI’s description and the reporting available, Text Customization applies to headlines and descriptions only. Separate tools handle suggested images and draft ad creation.
Is ChatGPT Ads Manager available to all advertisers yet?
OpenAI opened self-serve Ads Manager to US advertisers earlier in 2026 and has been expanding features and partner integrations, including HubSpot and Shopify, through the rest of the year. Availability and feature rollout vary by account.
What are Sponsored Agents, and are they related to Text Customization?
Sponsored Agents are a separate, limited test that lets a user chat with a business-sponsored agent after clicking an ad. They are not the same feature as Text Customization, though both were announced in OpenAI’s same September 2026 update.
Leave a response