Google sent us a legal notice. Here's what to do as a founder.

Arrr, ahoy Cap’n. Serpy here — the old parrot who’s watched a few storms roll through these waters. Reza spotted this one in the admin inbox, so I’ll ride along and point out the rocks. Grab a coffee; this one’s worth five minutes.

Google sent us a legal notice. Two, actually, two days apart.

We’re not in trouble. These are the notices Google sent to Google Workspace and Google Cloud admins, and ours landed in our admin inbox like everyone else’s.

But if you’re running AI agents on your Google accounts, the fine print now says whatever they do is on you. That part is worth five minutes.

What arrived

On 20 Oct our Workspace admin account got “[Legal Notice] Updates to Google Workspace Service Specific Terms and Agentic AI guidelines”. On 22 Oct the same inbox got “[Legal Notice] Updates to Google Cloud Terms of Service and AI guidelines”. Both say the changes took effect on 12 Oct 2026, and both include this line: “Your continued use of AI features is subject to these updated terms.”

We’re a small team that ships fast with AI, much like a lot of you. We use agents for code, drafts, research and inbox work. So I actually read the terms instead of archiving the email.

A wise captain reads the charts before the reef, not after. — Serpy

Does this apply to you?

If you pay for Google Workspace (your name@yourdomain email) or use Google Cloud, these are your terms now, whether or not you opened the email. If you didn’t see it, check the inbox of whoever set up your Workspace, since these notices go to admins.

If you’re on a free personal Gmail, these specific contracts aren’t the ones you agreed to. But Google’s general account rules still apply, and as you’ll see below, they now mention AI agents too.

What changed, in plain words

Five changes matter for founders running AI agents on Google accounts. Here they are at a glance, then in detail below.

Change	What it means for you	Where it’s written
Agent actions = your actions	Anything your AI agent does on Google counts as you doing it	Cloud ToS §3.3 (Restrictions)
You indemnify Google	If your agent gets Google sued, the bill can come to you	Cloud ToS §13.2
You’re solely responsible	Includes what the agent does and fails to do	Workspace Terms §13.3
A rulebook for agents	No click farming, no fake identity, no mass accounts	Agentic AI Responsible Use Guidelines
Tighter billing & termination	Faster suspension/termination; you pay your agents’ charges	Cloud ToS §2.1, §2.4, §8.3

1. Your agent’s actions count as your actions. The Google Cloud Terms of Service now name AI agents in the restrictions section. Section 3.3 says the customer will not “access or use the Services or any AI Agent … in violation of the AUP” (the Acceptable Use Policy). Source: Google Cloud Terms of Service.

2. You cover Google if your agent breaks the rules. Section 13.2 says the customer will defend and indemnify Google against third-party legal proceedings arising from “Customer’s or an End User’s use of the Services or AI Agents in breach of the AUP or Section 3.3 (Restrictions).” In plain terms, if your agent gets Google sued, that bill can come to you.

3. Workspace spells out who is responsible. The Workspace Service Specific Terms, section 13.3, say the customer is “solely responsible for: (i) any Agentic AI Action”. The definition of an Agentic AI Action includes “omissions or failures to execute such tasks”. So the agent forgetting to do something is also yours. Source: Google Workspace Service Specific Terms.

4. There’s now a rulebook for agents. Google published the Agentic AI Responsible Use Guidelines. The prohibited list includes “Fraudulently manipulating digital metrics through click farming, artificial engagement”, “obfuscating or misrepresenting agent identity”, and “Creating or managing numerous accounts to evade detection”. Google also “strongly recommends implementing human oversight and explicit approval mechanisms” before agents take consequential actions, and reminds you that “you remain liable for any charges accrued by your agents.”

5. Some billing and termination terms got tighter too. Google may suspend if you pay by a delayed method like a bank transfer and it “reasonably determines that Customer poses a risk of non-payment” (section 2.4). It can terminate immediately if it decides a material breach “is incapable of being cured” (section 8.3). And if it finds you used the service in a way meant to avoid fees, it may bill “for the entire month” (section 2.1).

Why a solo founder should care

When you’re one person, your Google account isn’t just email. It’s your domain email, your Drive, your login for half your other tools, and maybe your billing. Google’s own help page on disabled accounts now says: “Violations may result from instructions you give to an AI agent.” (Google Account Help).

Rocks ahead, Cap’n. The danger isn’t the lawsuit you’ll likely never see. It’s the agent that fires 300 cold emails in an afternoon — and your inbox goes dark the same day.

The 10-minute checklist

1. List every agent that touches your Google accounts. Coding agents with Gmail or Drive access, Zapier-style flows with an AI step, browser agents, “AI inbox” apps. If you can’t list them, start there.
2. Check scopes. Open your Google Account’s third-party connections and remove anything you don’t use. If an agent only needs to read, don’t give it send.
3. Put a human in front of anything that leaves the building. Emails, posts, replies, payments, deleting files. Drafts are cheap. Unsent mistakes are free.
4. Keep a log. What the agent did, when, and who approved it. If Google ever asks, you want an answer that isn’t “I’m not sure.”
5. No spam, no fake engagement, no posing as a human you’re not. All three are on Google’s prohibited lists.
6. Separate the blast radius. Give agents their own account where you can, so one bad run doesn’t take down the inbox your customers write to.
7. Set a spending cap. Google says you’re liable for what your agents spend.

This is my reading as a founder, not legal advice. Read the actual terms for your own setup.

The other side of AI agents

Here’s the part that’s harder to fix with a checklist. AI assistants aren’t only doing our work. They’re also how customers find products now. People ask ChatGPT or Google’s AI answers “what’s a good tool for X” and never see a list of links.

That’s why we built Serpair. It tracks whether Google and AI assistants like ChatGPT mention your site, and how that changes over time. If you want to see where you stand today, run the free check.

No captain should sail blind, Cap’n. If the AI’s naming your rivals and not you, you want to know today — not next quarter.

If you’re new to this, two of our guides pair well with the above: how to improve your AI visibility and AEO vs GEO vs SEO.

Frequently asked questions

What does the Google AI legal notice mean?

The Google AI legal notice means Google updated its Workspace and Cloud terms, effective 12 Oct 2026, to make customers responsible for what their AI agents do. If you received a “[Legal Notice] Updates to Google … Agentic AI guidelines” email, your continued use of AI features is now governed by these updated terms — your agent’s actions (and omissions) count as yours.

I received a legal notice from Google about AI — do I need to do anything?

You don’t need to reply, but you should act. Review which AI agents can access your Google accounts, limit their scopes, and put human approval in front of anything an agent sends or pays. The 10-minute checklist below is how to act on the Google AI notice without a lawyer.

How do I make sure I’m following Google’s AI agent guidelines?

Follow Google’s Agentic AI Responsible Use Guidelines: no click farming or artificial engagement, never misrepresent an agent as a human, don’t create accounts to evade detection, keep human oversight for consequential actions, and set spending caps. The checklist below turns each guideline into a concrete step.

Do Google’s new AI agent terms apply to a free personal Gmail account?

Not the Workspace or Cloud contracts — those bind paying customers. But Google’s general Account Help now states that violations “may result from instructions you give to an AI agent,” so a free Gmail is still at risk of suspension if an agent breaks the rules.

Who is liable if an AI agent breaks Google’s rules?

The customer is. Google Cloud ToS §13.2 makes you defend and indemnify Google against claims from your agent’s breach, and Workspace Terms §13.3 makes you “solely responsible” for every Agentic AI Action — including the agent’s omissions.

What counts as an “Agentic AI Action” under Google’s terms?

An Agentic AI Action is any task an AI agent performs on your behalf — and Google’s definition explicitly includes “omissions or failures to execute such tasks.” So an agent that forgets a step is still your responsibility.

What can get a founder’s Google account suspended for agent use?

Click farming or artificial engagement, obfuscating or misrepresenting an agent’s identity, and creating many accounts to evade detection — all on Google’s prohibited list. A single agent sending hundreds of cold emails can trip spam enforcement and disable the account the same day.

Who pays for what an AI agent spends on Google Cloud?

You do. Google’s Agentic AI Responsible Use Guidelines state plainly: “you remain liable for any charges accrued by your agents.” Set a spending cap.

Over to you

I’d really like to hear how other founders are handling this. How are you using AI in your business right now? Did Google notify you too? Have you changed anything since, and what guardrails do you have in place? Tell us in the comments or on Indie Hackers, and we’ll add the best guardrails to the checklist.