ChatGPT Ads: From October 30, AI Advertising Runs Without Your Approval if You Do Nothing

Imagine you place an ad for your company in ChatGPT. You have reviewed the copy, chosen the landing page and set the budget. A week later the ad is running in several variants, with different headlines, in a different language, and leads to a different subpage than intended. Nobody showed you the variants. This is not a thought experiment but the description of a default setting that, according to a trade report dated October 20, 2026, OpenAI will switch on for self-serve campaigns in ChatGPT Ads from October 30, 2026. That leaves seven days to decide for yourself.

For managing directors and marketing leads in mid-sized companies, this is a case study in a topic that becomes more important with every advertising platform: who decides what is published under your name? This article sets out what the sources say about the change, what we were able to check ourselves and what we were not, and how you can design your own approval step so that a platform default does not decide for you. A note on availability up front: the change is a default setting that applies from October 30. According to a trade publication, the self-serve platform itself has been running as a beta in Germany, Austria and Switzerland since August 31, 2026.

What is changing: "Creative Enhancements" in ChatGPT Ads

The source is a post by Barry Schwartz in the trade publication Search Engine Roundtable dated October 20, 2026, which we retrieved on October 23, 2026. It is based on an email in which OpenAI informed advertisers and on the associated OpenAI help page. An important note on the state of the sources: we could not retrieve the OpenAI help page itself; the server responded with an access block (HTTP 403). We therefore only report what the trade report quotes from the email and the help page. We have not read the original text at OpenAI ourselves.

According to this report, the following applies:

• What happens: OpenAI automatically enables "Creative Enhancements" for ChatGPT Ads, for new and existing self-serve campaigns.
• From when: according to the email, from Friday, October 30, 2026.
• What the feature can do: create new or adjusted headlines and descriptions, resize images, optimize the destination URL and translate ad copy into other languages.
• What it is based on: your existing ad and landing page content. According to the report, conversation context or user interests are taken into account. Variants can be created in advance or in real time.
• Approval: When enabled, according to the quoted help page, the ads can run automatically without you having to review or approve each individual version. The quote reads: "When enabled, these ads can run automatically without requiring you to review or approve each version."

The core of the announcement is the last point. What is new is less the ability to rewrite copy than the removal of individual approval as the default. If you do nothing, from October 30 variants will run that you have never seen.

Opt-out: what you can do and what the source says about it

According to the report, advertisers can opt out before October 30, 2026, or at any time, via the settings in Ads Manager. Existing opt-outs at account and campaign level are retained. Campaign-specific settings take precedence over account defaults; the email puts it this way: "campaign-specific settings will take priority over account defaults." In his post, the author also shows that the option is enabled in his own account settings.

What the source does not say is just as important. It does not describe where exactly in Ads Manager the setting can be found, and we have not clicked through it ourselves because we have no access to such an account. If you advertise in ChatGPT Ads, you should look at the setting at account level and at campaign level yourself before the deadline is reached. If you cannot find it there, contact OpenAI support.

Between "opt-in" and "opt-out" lies a decision. With opt-in, the platform has to convince you. With opt-out, you have to object to the platform, and anyone who does nothing has decided without noticing.

Reactions from practice: two voices, no more

The report quotes two reactions. One advertiser, Melissa Mackey, reportedly paused her ChatGPT Ads test because of poor lead quality. Another user, Anthony Higman, reportedly criticizes the presentation as a work simplification and alleges that the ads would also be served on unwanted topics. Both statements are individual voices from social media, which the author reproduces. They are not evidence that the feature works badly or well. We mention them because they show which concerns advertisers have: lead quality and the question of the environment in which their brand appears.

Availability: present in DACH, but still beta

For an assessment relevant to mid-sized companies, it matters whether the platform is usable for you at all. The trade publication onlinemarketing.de reported on August 31, 2026 that OpenAI's self-serve Ads Manager is opening for eligible advertisers in Germany, Austria, Switzerland and 28 other European markets. According to the report, these markets had received ChatGPT Ads one week earlier. Access starts as a beta version; advertisers create a beta account in Ads Manager for this purpose.

• Status of the platform: self-serve access as a beta (as of August 31, 2026, according to onlinemarketing.de).
• Status of the change: default setting for self-serve campaigns from October 30, 2026 (according to Search Engine Roundtable, October 20, 2026).
• Region: ChatGPT Ads, according to onlinemarketing.de, in Germany, Austria and Switzerland, among others.
• DACH: yes, with beta access. If you already advertise there, you are affected by the change.

This means: if you have a beta account in DACH and run campaigns, you are directly affected. If you do not advertise in ChatGPT, you do not need to change anything now, but you can take a question from this case and apply it to every other advertising platform: which AI features are already active at our company without our having deliberately switched them on?

What the change means for a mid-sized company

The reports say nothing about how good or bad the automatically generated variants are. We do not know, and we have not tested it. What we can assess is the structure of the risk, and it applies regardless of the quality of the copy.

First: the brand speaks without anyone having listened. An ad is public communication by your company. If software rewrites headlines, tone, promise or level of detail can change. A sentence that is legally sensitive in your industry, such as a promise of a cure or a price comparison, can arise this way without a person having seen it.

Second: the landing page can change. According to the report, the feature can optimize the destination URL. For your company, this means: the click may not land where you expected. This affects reporting, the privacy notices on the page and the question of whether the page fits the ad's promise at all.

Third: translations are your responsibility. If ad copy is automatically translated into other languages, it appears in markets whose language you may not be able to check. For a mid-sized company focused on the German-speaking region, this may sound unimportant. For an exporting business, it is not.

Fourth: control shifts from "before" to "after". Without individual approval, you can only establish after the fact what has run, for example in reports. That is less than a prior check, and it presupposes that you look at the reports regularly and know what to look for.

Opt-in or opt-out: why the default setting matters

Platforms can introduce new AI features in two ways. With opt-in, the feature is switched off until the advertiser deliberately switches it on. With opt-out, it is switched on, and the advertiser has to actively object in order to switch it off. Both routes are permissible and common. The difference lies in who bears the burden of the decision.

With opt-in, the platform carries the duty to deliver: it has to explain and convince. With opt-out, the advertiser carries it by omission: anyone who overlooked the notification has, under the platform's rules, agreed. For companies, a working habit follows from this that goes beyond ChatGPT Ads. Check the default settings of your advertising platforms not only at setup, but regularly and especially whenever the terms change. Designate a person who reads the platforms' notifications, and set a rhythm in which the settings are checked.

Whether such a default setting has legal consequences for liability or labeling, we cannot assess, and the sources say nothing about it. If you advertise in regulated industries, clarify this with your legal counsel before you rely on the platform default.

Which team at Amazing Agents is responsible: Daniel and Nadja

On amazing-agents.com, the topic belongs in the Research & Analytics department. Agent Daniel is responsible for analytics there. The website describes him this way: numbers without commentary are useless, Daniel makes data decidable. Agent Nadja heads the department and, according to the website, is there to ensure that content is based on research rather than opinion. The risk assessment of such a platform change, that is, the questions of which accounts are affected, which campaigns are running and how metrics change after the deadline, belongs in this task.

For the question of what is checked before publication, the website names agent Magdalena in the Tech & Ops department: no content goes live without her brand check. That is the idea we would adopt for advertising: an ad variant that appears in your name first goes through a check for brand and tone before it goes live, regardless of whether a person or software wrote it.

Here we describe an approach, not a result from a client project. In an audit, we would define with you which ad changes an agent may prepare, which samples Daniel evaluates and at which point a person approves before the ad is placed.

Agent with limits and an approval step: the difference

The ChatGPT Ads case is an illustrative example of what makes the difference between an agent without and one with an approval step.

Without an approval step: software rewrites ads, translates them and changes landing pages, and everything goes live immediately. You see the results in a report after the budget has been spent. If a variant is wrong, you have already paid for it and possibly shown your brand with it.

With an approval step: the agent prepares variants and submits them for review. A person checks tone, statements and landing page, and approves or rejects. Only then does the ad run. This costs time, and that time is the price of knowing what is being said in your name. An agent with limits and an approval step works quickly in preparation and leaves the decision with you.

On amazing-agents.com, this framework applies to all agents: approval steps for critical actions, and emails are created as drafts, not sent directly. What applies to emails can be transferred to advertisements: an ad is a draft until someone approves it.

What we do not know

• The original text: we could not retrieve the OpenAI help page. We rely on the trade report that quotes it.
• The quality of the variants: the sources say nothing about this, and we have not tested it.
• The exact location of the setting: where in Ads Manager the option is located is not stated in the sources we read.
• Legal consequences: whether labeling obligations or liability questions arise from this, we have not checked.
• Other platforms: whether other providers plan similar default settings, we cannot substantiate, and we draw no conclusions here.

What you should do this week: three steps

1. Check whether you are affected: if you have a ChatGPT Ads account, check whether "Creative Enhancements" is active in the settings at account and campaign level. Record the status in writing and with a date. If you do not advertise, you can skip this step.
2. Decide deliberately: decide before October 30 whether you want to use the feature and, if so, for which campaigns. You should not let a campaign with legally sensitive statements be rewritten automatically. Record the decision and the responsible person.
3. Define your own approval step: describe in one paragraph who sees an ad variant before it is placed, what is checked, for example tone, statements, landing page and language, and what happens if someone rejects a variant. Also define who checks the settings of all advertising accounts at regular intervals.

Conclusion

The change at ChatGPT Ads is less an isolated case than a pattern. Platforms build AI features into their advertising tools and configure them so that they run without any action on your part. For OpenAI, the sources establish: from October 30, 2026, "Creative Enhancements" are active by default for self-serve campaigns, ad variants can run without individual approval, and an opt-out is possible in Ads Manager. They do not establish how good the variants are, and we have not read the original text of the help page ourselves.

For you, this means: decide for yourself before the default setting decides for you. Define who approves. And remember the question for every advertising platform: what is already running automatically at our company without our having deliberately switched it on? An agent with limits and an approval step is a better foundation than trusting a default that someone else has chosen.

Sources: Barry Schwartz, "OpenAI Auto Enabling Creative Enhancements In ChatGPT Ads", Search Engine Roundtable, October 20, 2026 (retrieved on October 23, 2026; the OpenAI help page linked there could not be retrieved, HTTP 403). onlinemarketing.de, August 31, 2026 (availability in DACH, beta). Assignment to the agents: amazing-agents.com, home page and team pages (Daniel, Nadja, Magdalena; approval steps for critical actions).