Google's new ad terms put you on the hook for what its AI does

Google updated the Google Ads Terms of Service, effective 1 July 2026, to formally authorise automated program features to generate targets, ads, and destination URLs on the advertiser's behalf, and to clarify how inputs — including URLs and account access granted for automated campaign setup, and information entered into conversational campaign-building tools — can be used across Google's ad features to improve performance.

The update also restates, rather than removes, the advertiser's obligations: you are responsible for having the rights to whatever inputs you provide, and you are responsible for reviewing, approving, or removing any campaign or ad asset that Google's systems generate automatically. Google's own Help Centre page for the update states this obligation directly rather than leaving it to interpretation.

Feature launches are optional; terms of service are not. Every Google Ads account was bound to the new terms automatically on 1 July, which means the shift applies whether or not a given advertiser has actively enabled AI Max, Performance Max, or any conversational campaign-building tool.

The practical effect is that Google's automated systems now have explicit, standing authorisation to act on an account, and the advertiser's sign-off duty is the thing carrying the update, not the automation itself. Search Engine Land's coverage of the rollout frames it the same way: the terms describe automation as a system advertisers participate in, not a feature they select.

The advertiser is. The updated terms do not transfer liability to Google for automated decisions — they authorise Google to make those decisions and place the review burden on the account owner. If AI Max expands a search term into an irrelevant category, or Performance Max generates ad copy that misrepresents an offer, the terms are explicit that catching and correcting it is the advertiser's job, not something Google is contractually on the hook for undoing after the fact.

For B2B accounts, where a single misdirected campaign can burn a monthly budget on the wrong ICP or attach the wrong claim to a regulated offer, that review obligation is not a formality. It is the entire risk-management layer the account now runs on. This is a plain reading of Google's published terms rather than legal advice — where an offer carries real regulatory exposure, confirm the specifics with your own counsel.

• You are responsible for the rights to any input you feed into conversational or automated campaign tools.
• You retain a continued obligation to review, approve, or remove auto-generated campaigns and ad assets.
• Liability for what automation produces stays with the advertiser; Google's authorisation to act is not an assumption of responsibility for the output.

Treat every AI Max, Performance Max, or conversational-tool campaign as a draft until a human has reviewed the generated targets, ad copy, and destination URLs — the same discipline SalesHive applies before any client campaign goes live, regardless of which platform's automation built the first draft.

Beyond that single habit, three checks turn the new terms from a legal footnote into an operating routine.

• Audit what inputs (URLs, account access, conversational prompts) automated features can currently read, and confirm you hold the rights to all of them.
• Put a standing review step in the campaign workflow for anything AI Max or Performance Max generates, rather than trusting the automated output by default.
• Document who signed off on each auto-generated asset — the terms make that review your obligation, so the record of having done it is now part of account hygiene, not just good practice.