Google Rewrote What Every Advertiser Agreed To. Nobody Had to Click Anything.

On 1 July 2026, every active Google Ads account became subject to a revised Terms of Service. There was no login prompt. No banner requiring a click to continue. No email requiring a reply. The update simply applied, silently, across the platform.

The substance of the change is more consequential than its quiet rollout suggests. The revised terms state that advertisers "explicitly authorize" 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." Advertisers remain responsible for reviewing, approving, editing or removing any campaign or ad asset - including those Google's own tools generate automatically - but the starting position, the contractual default, has moved.

From feature to default

To understand why this matters, it's worth being precise about what changed. Google's automated ad-generation tools - Performance Max, Smart Bidding, automatically-created assets, AI Max - are not new products introduced by this terms update. Versions of most of these tools have existed for several years, and advertisers have always been able to opt into using them.

What changed on 1 July was not the existence of the tools. It was the legal baseline governing their use. Previously, automation was something an advertiser (or the agency managing their account) actively chose to enable for a given campaign. Under the revised terms, every advertiser has, by virtue of using Google Ads at all, already granted standing authorisation for Google to apply automated features - formatting, targeting, ad generation, destination selection - without a separate, campaign-specific opt-in.

The distinction between "a feature you can choose to use" and "a standing authorisation you've already granted" is not merely semantic. It shifts where the burden of active decision-making sits. Under the old model, an agency or advertiser had to actively decide to hand a specific task to Google's automation. Under the new model, Google's automation has standing permission by default, and it is the advertiser's responsibility to actively notice, review and, if necessary, override whatever the automation has done.

Who carries the responsibility - and who doesn't

The revised terms are explicit that advertisers remain legally responsible for reviewing and approving AI-generated campaign elements. This is a meaningful detail: Google has structured the change so that expanded automation authority comes with no corresponding expansion of Google's own liability for what that automation produces. If an automatically-generated ad, target audience, or landing page destination causes a problem - a brand safety issue, a compliance breach, a factually incorrect claim - the responsibility for having failed to catch it sits with the advertiser (and, in practice, with whichever agency is managing that advertiser's account), not with the platform that generated it by default.

This creates an asymmetry worth naming clearly. Google gains broader standing authority to apply automation across every account, with no advertiser needing to take any action to grant that authority. Agencies and advertisers gain an expanded review burden - a wider set of AI-generated decisions they now need to actively check, on top of whatever they were already reviewing - with no reduction in their existing liability if something is missed.

Why this specifically threatens agency billable work

For marketing agencies, the practical implication runs deeper than a compliance footnote. A meaningful share of agency billable work in paid media has historically been built around exactly the tasks this terms update now defaults to automation: formatting ad assets for different placements, selecting audience targets, generating destination URLs and landing page variants based on campaign objectives. These were traditionally functions an agency's media buyers and account managers performed manually, or at minimum actively configured and monitored.

When the platform itself has standing contractual authority to perform these functions automatically, the agency's traditional value proposition in that specific layer of the work - "we build and configure your campaign so it runs correctly" - becomes harder to distinguish from what the platform does unprompted. The agency's job doesn't disappear, but it shifts: from actively building campaign elements to actively auditing campaign elements the platform has already built, catching errors, and explaining discrepancies to clients who may not realise their campaign changed itself.

The disclosure gap this creates

This terms change lands alongside a separate but related Google announcement: a new "How this ad was made" disclosure feature inside My Ad Center, rolled out 9 July 2026, letting any consumer check via a three-dot menu whether an ad was built using Google's generative AI tools. Disclosure triggers automatically when Google's own AI tools are used. If an ad was produced elsewhere - by an agency, using non-Google tools - the advertiser must manually flag that AI was involved. Google does not independently verify this.

Read together, the two changes create a specific gap that agencies now need to actively manage: campaign elements that Google's own automation quietly generates under the new default-authorisation terms may not be clearly flagged to the client as AI-generated, while the disclosure burden for anything produced outside Google's tools sits entirely with the advertiser (and by extension, the agency) to self-report accurately.

What agencies should actually check

For any agency running client campaigns through Google Ads, the practical response to this shift is straightforward, even if it requires establishing a new habit rather than a one-off fix. First, audit which automated features are currently active on each client account - Performance Max, automatically-created assets, Smart Bidding variants - since many of these may already be running by default rather than by deliberate configuration. Second, establish a recurring review cadence specifically for AI-generated campaign elements, separate from standard performance reporting, since the terms update means these elements can appear or change without any notification. Third, build client communication around this explicitly: clients should understand that some of what appears in their campaign may originate from Google's automation rather than the agency's direct work, and that the agency's value now includes actively catching and correcting what the automation produces.

The agencies that will hold client trust through this shift are not the ones who simply keep doing what they did before and hope nothing changes underneath them. They are the ones who can clearly articulate, to a client who asks, exactly which parts of the campaign are the agency's direct work, which parts are the platform's automation, and who is actively checking the latter.