On July 1, 2026, Google Ads updated its terms of service. No pop-up, no "I agree" checkbox, no email confirming you'd read anything. Every account was simply bound to language that authorizes Google and its affiliates to "format, select, or generate targets, ads, or destinations" using automated features, by default. The same week, the 2026 CMO Survey put a number on the other half of the story: marketing budgets grew just 1.3% year over year, and the money that did move went toward paid media, now 31.4% of the average budget, funded by cuts to agency spend. The system writing more of your ads just got more authority. The humans who'd catch a mistake before it went live just got fewer.

Layer in the regulatory backdrop and the timing gets sharper. Google's ad-market monopoly case is moving into its remedies phase, the point where courts decide how the open web's ad pipes get rebuilt, right as the AI-in-advertising tools market is projected to hit $14.12 billion this year alone. Google is under more scrutiny for how it controls the auction and handing advertisers more automated control inside that same auction at the same time. Those two facts are not in tension. They're the same strategy: keep the volume high and the liability downstream.

Why didn't Google ask for a signature?

Because it didn't need one: automation stopped being an optional program feature and became the default account behavior, so there was nothing left to opt into. The new terms are explicit that advertisers remain "responsible for reviewing, approving, editing, or removing" any campaign or asset the system generates on their behalf. Production moved to the machine. Liability never moved anywhere.

What does the agency cut actually cost?

It costs more than the line item saved, because what gets cut isn't a fee, it's a layer of scrutiny. PwC's study of 22,000 respondents across 32 countries found that trust now shapes 92% of purchase decisions, and brands in the top quartile of trust indices earn a 31% revenue premium over lower-trust competitors. Weigh that against the few percentage points saved by trimming an agency of record, and the math doesn't favor the cut. It gets worse on the team side: the share of CMOs who feel their teams are equipped to handle what's coming dropped from 52% in 2025 to 42% in 2026. Budgets loosened slightly. Review capacity didn't.

Why does a human sign off without actually checking?

Because automation research has a name for exactly this behavior: the "moral buffer" effect. Once a system is perceived to be in charge, the human operator's sense of personal accountability drops, and approval becomes a reflex instead of a review. That's what automation bias measures, the tendency to treat a system's output as a shortcut past vigilant checking rather than an input to it. A one-click approve button is not a review process. An actual review requires understanding why the system surfaced that specific ad, and that takes time nobody budgeted for once the reviewers were the ones getting cut.

• Put a weekly "why this ad" review on every automated campaign. Ask for the reasoning behind the output, not just a yes or no.
• Fund oversight before you fund volume. If the agency line gets cut, redirect part of it to one named person accountable for catching automation errors before launch.
• Price the brand-safety miss into the plan. One viral mistake won't erase the full 31% trust premium, but it will eat a meaningful slice of it, and that number belongs in the budget conversation, not just the crisis one.

The AI writing the ad doesn't make the liability disappear. It just makes it harder to see whose name is on it.

What this means for your next quarter

Restoring the agency budget isn't the point, somebody still has to own the review. The practical move is a standing weekly check on every automated campaign that asks for the system's reasoning, not just a rubber stamp. With the AI ad tools market climbing toward $14.12 billion this year, everyone has access to roughly the same automation. What separates you from a competitor isn't which tool you licensed. It's who actually reads what the tool produced, and how carefully, before it goes live. Google changed its contract quietly. Your review process needs the same update, or the next mistake gets billed to your brand's trust, not your ad budget.