New Google Ads Terms Make AI Mistakes Your Fault

If you manage your own Google Ads account, you have likely received a notification regarding updates to the Google Ads terms of service taking effect on 1 July 2026.

While legal updates are traditionally ignored, this specific change requires your immediate attention.

Google is expanding its automated systems. They are clarifying exactly how your data fuels these systems and, crucially, establishing who carries the legal liability when the automation inevitably makes an error.

Here is the exact reality of what is changing, who it impacts, and the specific steps you must take to protect your campaigns and your business.

What is changing in Google terms of service?

The update explicitly defines the boundaries of advertiser responsibility as machine learning (AI) takes a more prominent role in campaign creation.

In reality, Google is formally stating that while they provide the automated systems to build and serve your adverts, you are legally and operationally accountable for the final output. You cannot simply set an automated campaign live and assume the platform understands the nuances of your industry.

Who is responsible for AI-generated ads on Google?

You are. The updated terms reinforce the necessity of a “human in the loop”.

Even if the platform’s machine learning builds your headlines, generates video variations, or selects your keywords, you remain entirely responsible for reviewing and approving those assets. The Google Ads policies and human review guidelines dictate that any misrepresentation, whether manually typed by you or dynamically generated by a machine, falls entirely on your shoulders.

How Your Data Fuels Google Website AI

When you input your URLs, copy, and product details into the platform, that data acts as the immediate fuel for their generative systems.

The conversational experience in Google Ads allows advertisers to rapidly build campaigns by simply providing a landing page and a brief description. The system crawls this data to understand how to better match similar businesses with potential customers. Furthermore, as Google prepares to test ads in generative AI search results, the reliance on your provided data to formulate accurate answers for searchers will only increase.

What happens if a Google Ads automated asset makes a false claim?

This is where you must understand the context of the platform. Think of Google Ads like a self-driving car. The platform provides the route to your destination, the engine and the seats and controls speed, braking and acceleration. You are still the driver.

You still have to stop it driving off a cliff in error.

If you leave the AI completely unchecked, you risk Google AI ad errors. If a machine-generated headline inadvertently makes a false claim, you are immediately in breach of the misrepresentation policy of Google Ads. This constitutes false advertising on Google Ads, which can result in account suspensions. You cannot blame the car if you refuse to look out of the window to see what’s coming.

Real-World Scenarios: What This Means In Reality

To clearly understand how these policies translate into daily campaign management, we need to look at specific situational examples.

Here is how automation and human oversight must work together to deliver what we want to deliver:

• Scenario A: The Seasonal Bakery vs. The Conversational Experience
  • The Setup: A local bakery uses the conversational experience to build a summer campaign for a “lemon drizzle loaf”.
  • The Reality: The system uses those inputs to build the ad copy dynamically.
  • The Risk: If the system incorrectly labels the loaf as “Gluten-Free” when it contains flour, the bakery owner must catch and correct this immediately. Failing to do so results in severe customer liability issues.
• Scenario B: The Local Plumber vs. Final URL Expansion
  • The Setup: What is final URL expansion? It is a feature where Google ignores your specific provided link and instead crawls your entire site to find the most relevant landing page for a user’s search.
  • The Reality: By enabling Google Ads final URL expansion, you authorise the platform to use any crawled content to dynamically build adverts. What does final URL mean in this context? It means the specific page the user eventually lands on, chosen by the machine.
  • The Risk: If the plumber updates their website to remove a specific service—like “Underfloor Heating”—they must ensure their ad assets are updated. If the system continues to generate adverts based on outdated cached content, the plumber pays for clicks they cannot service.
• Scenario C: The Clothing Brand vs. Automatically Created Assets
  • The Setup: A retailer utilises Google Ads’ generative AI for responsive search ads, uploading static photos while the system generates headlines.
  • The Reality: The system tracks Google AI responsive search ads updates to continually test new combinations of text and imagery.
  • The Risk: If the system unintentionally combines a photo of a summer dress with a headline about a “Winter Sale”, the retailer must identify and remove that specific combination. Where can I find the asset library in Google Ads? You must navigate to your campaign settings, select ‘Assets’, filter by ‘Added by’, and select ‘Google AI’. How to identify AI ads? They will be tagged directly in this filtered view.

Dispute Resolution and Account Legalities

The impending changes also formalise dispute resolution mechanics. The platform is introducing Google Ads terms of service arbitration rules, specifically “batch arbitration”. This allows them to group similar legal claims together for efficiency.

Are AI adverts legal?

Yes, provided they adhere strictly to local advertising standards.

Can I get a refund for Google Ads if the machine makes a mistake?

Generally, no. Because the terms state that you are the final point of approval, the financial liability for wasted clicks rests with you.

How to Protect Your Account: A Step-by-Step Action Plan

So, what to do if Google AI is wrong? You must take control, and adjust the mechanics.

Here is exactly what you need to do to fix the actual problems:

Step 1: Audit Your Automated Settings

• Navigate into your account settings and locate your asset controls.
• You must understand how to turn off automatically created assets in Google Ads if they are consistently hallucinating details about your business.
• If the data is constantly incorrect, turn off automatically created assets in Google ads entirely and revert to manual control. How do I turn off Google AI assets in Google Ads? Go to Campaign Settings > Automatically Created Assets > Select ‘Off’.

Step 2: Master Performance Max Configuration

• Performance Max campaigns rely heavily on machine learning to drive conversions across Search, Display, YouTube, and Gmail.
• They use your provided assets and combines them dynamically to create ads
• You should only use Performance Max campaigns when you have clear, accurate conversion tracking and high-quality creative assets.
• To run effective Performance Max campaigns, you must feed it high-quality conversion data (and lots of it, ideally). Poor data leads to poor Google Ads asset optimisation and poor campaign performance.
• To optimise Performance Max campaigns, you should exclude irrelevant URLs, refine your audience signals, and strictly monitor your asset performance grouping.

Step 3: Conduct Contextual Health Checks

• You cannot look at data in isolation. If your conversion rate drops, do not immediately blame the machine.
• Consider the environmental factors. Was it a bank holiday? Was there a major sporting event? You must look at the data in the context of the wider market.

The Bottom Line: AI Can Make Mistakes

Machine learning in digital advertising is designed to process vast amounts of data quickly, efficiently achieving the goal of connecting you with buyers. However, the machine does not understand the nuances of your specific commercial reality.

AI can make mistakes. You must monitor what it is doing closely to save your budget and avoid potential legal claims against your business.

Stay vigilant. Audit your assets regularly, and make sure a competent human checks on machine output.