On 1 July 2026 a revised Google Ads Terms of Service took effect across live advertiser accounts, and Search Engine Land reported on 2 June 2026 that no advertiser action was required before the rollout. Two weeks later, on 13 July 2026, Search Engine Journal covered a second change moving in the same direction: a new control for disclosing AI-generated ad creative, rolling out through July. Taken together, these are not cosmetic. They redraw the default relationship between a small advertiser and the machine spending their money. All sources below were opened and checked on 26 July 2026.
What actually changed in the Google Ads terms on 1 July 2026?
The same Search Engine Land report by Anu Adegbola, published 2 June 2026, describes the previous position differently: Google could provide tools to help advertisers generate targets, ads or landing destinations, with opt-in and opt-out options across many automation features. The article also notes the update applies to Google Ads accounts only, not Google Workspace or Cloud Identity, and that the terms expand how advertiser inputs may be used across AI features, including inputs typed into conversational tools.
Read the clause slowly and three verbs do the work. Format covers reshaping assets you supplied. Select covers choosing targets and placements. Generate covers creating new targets, ad variations or landing destinations that never existed in your account.
Why does one sentence in a legal document matter to a small advertiser?
For a six-person plumbing firm in Leeds or a two-chair salon in Cardiff, the practical question is not legal exposure to Google. It is what appears on screen when a customer searches. Automatically generated headlines, sitelinks and destinations are written from your website content and account signals. If your website still lists a service you dropped in 2024, an emergency callout promise you no longer honour, or an old price, the generation step can surface it as a live advert.
Search Engine Land's report is explicit that advertisers must ensure they hold the necessary rights to all information they provide, and must continue reviewing, approving, editing or removing automatically generated campaigns and assets. The obligation to check did not move. What moved is how much there is to check.
Does the AI labelling change apply in the UK?
The Google Ads advertising policies update on AI labelling requirements confirms the mechanics. Starting in July 2026, advertisers may add text or visual labels within image and video ad creatives that were generated or modified using AI, and those labels will not breach Google's usual policies against text overlays and watermarks. Google's own tools may apply labels automatically. The rollout is gradual through the month across Google Ads, Display and Video 360, Campaign Manager 360, Merchant Center and Ads Editor.
One line in that page deserves highlighting. Google states that using the AI label setting does not guarantee compliance with specific regulations, and advises advertisers to seek legal guidance and take measures as needed so ads follow local obligations. The label is a disclosure mechanism, not a legal defence.
What is the difference between Google's own AI and third-party AI here?
Search Engine Journal's coverage by Brooke Osmundson, published 13 July 2026, reports that advertisers using third-party generative AI tools must manually disclose AI use, while Google's tools apply the label themselves. It quotes Google saying that when advertisers create ads elsewhere, a control is being introduced so they can easily indicate if they used generative AI.
That report also notes the labels surface to the public through a section in My Ad Center called "How this ad was made", reachable from the three-dot menu or info icon on ads across Search, YouTube and Discover. Osmundson's practical point is that organisations may need to update approval processes to document AI usage before campaigns launch. For a small firm using a freelancer or a small agency, that means a simple question added to the handover: which parts of this creative came out of a generative tool.
What do the UK advertising rules actually say about disclosing AI?
The ASA guidance published on 29 May 2025 sets out a two-question test for advertisers. Is the audience likely to be misled if the use of AI is not disclosed. If there is a danger of the audience being misled, is disclosure clarifying the ad's message or contradicting it.
The same guidance adds a caution worth memorising: disclosure alone is very unlikely to mitigate the harm caused by a fundamentally misleading message. Its worked example is an AI-generated image falsely showing the effects of a cosmetic product, which no label can rescue. That framing maps cleanly onto Google's own wording that the label setting does not guarantee regulatory compliance. Two different bodies, same conclusion. Labels describe how something was made, not whether it is true.
Which settings should a small advertiser check this week?
Automatically created assets is the setting that lets Google write headlines and descriptions from your site. It can be switched off at campaign level, and turning it off is a reasonable choice for regulated trades, anything with pricing complexity, and any business where a wrong claim carries real cost.
Final URL expansion allows Google to send traffic to pages other than the one you nominated. On a site with thin or outdated pages, that is where irrelevant landings come from. Asset reporting matters because it is the only place you see what was actually written on your behalf. Your website matters because generation reads from it, so an out-of-date services page becomes an out-of-date advert. Budget planning shifts too, which is worth reading alongside a broader UK marketing budget breakdown for 2026 before reallocating spend.
What are the trade-offs of accepting more automation?
Google's own framing in the terms is that advertisers keep the right to edit or remove what is generated. That right is only worth something if someone exercises it. A restricted setup that nobody reviews is not safer than an automated one that gets checked weekly. The variable is attention, not the toggle.
How does this connect to the wider shift towards AI-written marketing?
Search Engine Journal's report notes the disclosure control exists because AI regulations in the EU, India and New York require certain AI-generated or edited assets to carry disclosures. Regulation is running ahead of habit. Most small advertisers have no documented record of which assets were machine-made, which is exactly the record the new control assumes you can produce. The same question is being asked about editorial content, covered in more depth in this analysis of whether AI content hurts SEO in 2026.
The connective tissue is provenance. A business that can say where each claim came from handles both changes easily. A business that cannot is exposed twice.
What should a UK owner do in the next 30 days?
First, open every active campaign and record whether automatically created assets and final URL expansion are on. Record it, do not just glance at it, because the point is having a baseline you can compare against next quarter.
Second, read your own site as if a machine were quoting it. Every service listed, every response time promised, every price shown, every accreditation claimed. Anything you cannot evidence today should be removed today, because generation reads from it.
Third, set a recurring 20-minute review of the asset reports, and add one question to any brief you hand a freelancer or agency: was generative AI used for any part of this creative, and where. That single question is what makes the new disclosure control usable rather than theoretical, and it is the kind of control that belongs in any serious paid media setup.
The pattern behind both changes is the same. Google is widening what its systems may do by default, and narrowing nothing about who answers for the result. For a UK small business, that is not a reason to abandon the platform. It is a reason to treat the account as something that needs reading, not just funding.
Google Ads AI Terms 2026 — FAQ
Did the new Google Ads terms need me to accept anything?
No. Search Engine Land reported on 2 June 2026 that the revised Google Ads Terms of Service took effect on 1 July 2026 and that no advertiser action was required before the rollout. That is the part owners keep missing. There was no consent screen, no email you had to click, no settings migration. The terms simply applied to live accounts on the date. The practical consequence is that if you assumed nothing changed because nothing appeared in your inbox, your assumption was wrong. The sensible response is a settings audit rather than a legal review, because the change is about what Google is authorised to do by default inside campaigns you already run.
Does the AI ad labelling rule apply to businesses in the UK?
Not as a Google-enforced legal mandate. The Google Ads policy help page on AI labelling requirements names the European Union, India and New York as the places where regulations require disclosure for ads containing certain AI-generated or edited assets. The United Kingdom is not on that list. The label setting itself is rolling out across Google Ads, Display and Video 360, Campaign Manager 360, Merchant Center and Ads Editor from July 2026, so UK advertisers will see the control appear regardless. Google also states plainly that using the AI label setting does not guarantee compliance with any specific regulation, which means the control is a tool, not a shield.
Who is responsible if Google's AI generates an inaccurate ad for my business?
The advertiser. Search Engine Land's coverage of the terms update states that advertisers must ensure they hold the necessary rights to all information they provide, and must continue reviewing, approving, editing or removing automatically generated campaigns and assets. That responsibility does not shrink because the asset was machine-made. For a UK trade business, the risk is concrete: an automatically generated headline claiming a guarantee you do not offer, a service area you do not cover, or a price you no longer charge. The cheapest control is a weekly look at the automatically created assets report and a habit of removing anything you would not say yourself.
What does the new automation clause actually say?
Search Engine Land quotes the new wording as: Customer authorizes Google and its affiliates to serve ads, including through the use of automated program features to format, select, or generate targets, ads, or destinations on Customer's behalf. The same report describes the previous position as Google providing tools to help advertisers generate targets, ads or landing destinations, with opt-in and opt-out choices around many automation features. The shift is from tools offered to authority granted. Three verbs carry the weight. Format means reshaping what you supply, select means choosing targets and placements, and generate means creating new assets, targets or destinations that you did not write.
How do UK advertising rules treat AI-made ads?
The Advertising Standards Authority published guidance on 29 May 2025 confirming there is no blanket legal requirement in the UK to disclose the use of AI in ads, and that existing rules apply regardless of how content is generated, edited or targeted. It offers a two-question test: is the audience likely to be misled if AI use is not disclosed, and if so, does disclosure clarify the message or contradict it. The ASA also warns that disclosure alone is very unlikely to mitigate the harm caused by a fundamentally misleading message. So a truthful AI-assisted ad needs no UK label, and a misleading one is not rescued by adding one.
