TL;DR
- Meta's revised Commercial Terms take effect on 30 October 2026 and add a new AI Features section that is not in the current version, Adweek reported on 22 October.
- Advertisers become 'solely responsible' for output of Meta AI features, including features Meta switches on; the terms say the AI 'may make mistakes' and put 'any and all risks' on the advertiser.
- The earlier generative AI ad terms only applied when an advertiser chose to use the Ad Creative AIs 'in your sole discretion'.
- Meta says individual Advantage+ features can be turned off in Ads Manager, but did not directly say whether advertisers are liable for features they never enabled.
Meta's revised Commercial Terms, which take effect on 30 October 2026, add a new "AI Features" section that makes advertisers "solely responsible" for whatever Meta's AI features create or change in their ads, including features Meta itself switches on. Adweek reported the clause on 22 October 2026 after reading the updated terms, and noted that the section does not exist in the current version of the document. Any business that keeps advertising on Facebook, Instagram or other Meta products after 30 October accepts it.
A different clause from the crawling and AI training story
Relevant Audience covered the same 30 October Commercial Terms twice already, on 17 October and 20 October 2026. Those posts dealt with the clause that lets Meta send crawlers to business websites, and with Meta's later removal of the word "train" from that crawling clause. The AI Features section is a separate part of the same document. It is not about what Meta collects from your website. It is about who carries the legal risk when Meta's own generative tools add to, rewrite or alter an ad.
Meta's preview of the revised terms now shows the crawling clause reworded, without the word "train", and Adweek's 22 October report quotes the new wording. According to the outlet, businesses authorise Meta to send "web scrapers, bots, robots, spiders, crawlers, and user agents" to sites associated with their products and services, and Meta may use what it collects to "provide, develop, improve, research, and maintain advertising and business products." A Meta spokesperson told Adweek the crawling terms "only applied to our ads and business products" and that content owners can still use robots.txt to signal their preferences. That thread is now a separate issue. This post is about the AI Features section.
What the AI Features section says
Adweek's reporter, Trishla Ostwal, read the revised terms and quoted three passages. Adweek's report does not reproduce the full section, but Meta's own preview of the revised terms, dated effective 30 October 2026, does, and the passages Adweek quoted match it.
- Advertisers will be "solely responsible" for whatever Meta's AI features create or change, and Adweek says this reaches features Meta switches on, not only those an advertiser chooses.
- Meta concedes in the new terms that its AI features "may make mistakes and may not always produce accurate, complete, or suitable results."
- The terms go on to say that "any and all risks" are on the advertiser.
The update also carries the usual exit. "If you do not agree to the updated terms, you may stop all access or use of our Products for business or commercial purposes," the update reads, according to Adweek. Meta's terms add that continuing any business use of its products after notice of an update means agreeing to be bound by it, so there is no signature or checkbox. Running ads after 30 October is acceptance.
How this differs from Meta's earlier generative AI ad terms
Advertisers already carried liability for AI-generated ad output before this update, but only when they chose to use the tools. Adweek quoted the earlier wording, which said advertisers "may choose to use the Ad Creative AIs in your sole discretion" and accepted that the output "may be inaccurate, incomplete, misleading, offensive, and/or inappropriate." The same text said the advertiser, not Meta, is responsible for its use of any output.
The change sits in the trigger. Under the old wording, responsibility followed the advertiser's choice to use a tool. Under the new section, as Adweek reads it, responsibility attaches to the AI output whether the advertiser switched the feature on or Meta did.
The table below sets the two states side by side, using what Adweek reported and Meta's published preview of the revised terms.
| Point | Earlier generative AI ad terms | Commercial Terms from 30 October 2026 |
|---|---|---|
| When the advertiser is responsible | When it chooses to use the Ad Creative AIs "in your sole discretion" | "Solely responsible" for AI output, including features Meta switches on |
| Meta's statement on accuracy | Output "may be inaccurate, incomplete, misleading, offensive, and/or inappropriate" | AI features "may make mistakes and may not always produce accurate, complete, or suitable results" |
| Where the risk sits | Advertiser responsible for its use of any Output | "Any and all risks" on the advertiser |
| How it is accepted | Choosing to use the tools | Continuing to use Meta products for business after 30 October |
What Meta said, and what it did not say
A Meta spokesperson told Adweek that the company is folding its existing generative AI ad terms into its main ad terms, which frames the change as consolidation. The same spokesperson said advertisers can turn off individual Advantage+ features in Ads Manager at any time.
Three things were not answered. Meta did not directly address whether the new terms make advertisers liable for AI features they never switched on. It did not comment directly on the AI Features section itself. And, according to Adweek, it did not respond to multiple specific questions about advertisers' experiences with its AI features. Meta's own preview text does not limit the section to features the advertiser enabled, so until Meta clarifies it, the safest reading for an advertiser is the plain one Adweek gives: if the AI changed the ad, the advertiser owns the result.
What advertisers say is already happening to live ads
Most of Adweek's report is about live campaigns rather than contract wording. Buyers described AI changes they say are already showing up in their ads.
Bookshop.org
The online bookseller Bookshop.org said in a post on Threads that an automated account update from Meta opted it into a generative AI feature, which added an AI-generated illustration to a human-made ad without its knowledge. The retailer wrote that the ad was "enhanced" or, in its words, "ruined, really" after the update, and it has since turned off all of Meta's AI ad features.
Sitelinks, translations and AI images
A digital agency told Adweek that Meta's AI is adding sitelinks, translating copy and generating AI images and carousels from a brand's own photos, sometimes on ads where the advertiser did not switch the feature on. A specialist there said Meta auto-applies the enhancements and that "many of the options are tucked away in hidden or collapsed menus." In the same specialist's words, "If every one of those isn't manually unchecked, the enhancements still apply." Even after several rounds of quality checks, that team has had to pull live ads to strip out changes that slipped through.
Buyers switching automation off
Another buyer quoted by Adweek said its team systematically turns off Meta's default AI automation on every campaign, to keep ads on the requested audience and geography and to avoid unwanted changes and legal exposure. A third agency head said Meta is opting advertisers into AI features a month before Black Friday and Cyber Monday, and that his firm is having its legal team review the terms but will not tell clients to pull back in the fourth quarter.
Why the timing matters
The terms take effect on 30 October 2026, which puts the change at the start of the year's heaviest retail period for many advertisers. Adweek also placed it in a busy week for Meta: on 21 October Meta reported US$64.6 billion in third-quarter revenue, and a New Mexico judge ordered the company to pay US$25 billion after a jury found that Facebook misled users about control over their data. Meta is expected to appeal that ruling.
PPC Land, which compared Adweek's reading against the terms currently in force (dated 6 March 2024), noted that the existing document already caps Meta's liability at the greater of US$100 or the amount a business paid in the previous twelve months, and already requires businesses to indemnify Meta against third-party claims related to their content. Read together with the new section, the practical effect is that an AI-made error in an ad is treated, contractually, as the advertiser's error.
How to audit Advantage+ creative enhancements before 30 October
The only control Meta pointed to is the ability to turn off individual Advantage+ features in Ads Manager. The checklist below is Relevant Audience's analysis of how to use that control in the days before the terms change. It is not a Meta procedure, and Meta's menus can differ by objective, format and account.
- List every live and scheduled ad. Include ads in always-on campaigns, catalog ads and anything scheduled to start after 30 October. Old ads are the easiest to forget.
- Open the creative settings at ad level. Advantage+ creative options are set per ad. Expand every collapsed group in that section, since one buyer told Adweek the options sit in hidden or collapsed menus and apply unless each one is unchecked.
- Decide feature by feature. Adweek's sources named added sitelinks, automatic translation, AI-generated images and carousels built from brand photos. For each enhancement, decide whether you are prepared to be "solely responsible" for what it produces. If not, switch it off.
- Check after account updates. Bookshop.org said an automated account update opted it in. Re-check settings after any notice from Meta about new ad features, not only when you launch a campaign.
- Keep dated evidence. Screenshot the settings for each ad with the date. If a complaint arrives later, you will want a record of what you allowed.
- Review delivered variants where you can. Look at ad previews and any reporting that shows which versions ran. Where you cannot see what a user received, treat that ad as higher risk.
- Brief whoever builds the ads. If an agency or freelancer manages your account, agree in writing who checks these settings and how often.
What this means for Thai marketers
Meta's Commercial Terms govern every business using its products, so the 30 October update applies to advertisers in Thailand as it does elsewhere. Adweek did not report any country exception. This section is analysis.
The exposure is highest where wording is regulated. Thai advertisers in health products, cosmetics, supplements, alcohol and financial services already work under strict rules on claims. An automatic translation that changes a claim, an added sitelink to the wrong page, or an AI image that shows a result the product cannot deliver is the kind of change the new section places on the advertiser. Bilingual campaigns deserve extra attention, because automatic translation of Thai or English copy was one of the enhancements buyers described. Teams running Facebook Ads through the Q4 sales period should complete the audit above before 30 October, and anyone planning wider paid social campaigns should add an AI-settings check to the launch routine.
FAQ: Meta's AI Features clause
When do Meta's new AI Features terms start?
They take effect on 30 October 2026, as part of the revised Meta Commercial Terms. Continuing to use Meta products for business after that date counts as accepting them, under the Updates clause in Meta's Commercial Terms.
Am I responsible for AI changes I never switched on?
On Adweek's reading, yes: the section makes advertisers "solely responsible" for AI output, including features Meta switches on. Meta did not directly answer that question when Adweek asked, so the full legal position is not stated by the source.
Can I turn Meta's AI ad features off?
Yes, Meta told Adweek that advertisers can turn off individual Advantage+ features in Ads Manager at any time. Buyers quoted in the report warn that some options sit in collapsed menus, so each one has to be checked.
Is this the same as the crawling clause reported earlier in October?
No, it is a separate section of the same document. The crawling clause covers Meta's bots visiting business websites; the AI Features section covers responsibility for what Meta's AI creates or changes in ads.
Does this apply to advertisers in Thailand?
The source did not report any country exception, and the Commercial Terms govern every business using Meta's products. Thai advertisers should assume it applies to them from 30 October.
The bottom line
From 30 October 2026, an AI-generated or AI-altered Meta ad that misleads or offends is the advertiser's problem, even when Meta turned the feature on. The work is not complicated, but it has to be done ad by ad and repeated after account updates. If you want a second pair of eyes on your Advantage+ settings before the deadline, Relevant Audience's paid social team can review them with you.
