Product 24 min read

Before You Connect Your Files to an AI Assistant, Read Its Ads Sentence

MMNMNOTE
AI privacypersonalized adsprivacy policyAI assistantsconnected apps

From at least July 2025 until October 2026, Google's privacy policy said it does not show personalized ads based on your content from Drive, Gmail, or Photos. Since October 16, 2026, that sentence ends with "unless you choose." Before you connect your files to an AI assistant, find its version of that sentence — every provider has written one.


Every quote below was read from the provider's own page on October 25, 2026, and each page's own date is printed beside it. Policies change, and they vary by region. This is a reading of published text, not legal advice.


The before is one flat line. Google's privacy policy, in the versions effective July 1, 2025, May 26, 2026, and October 1, 2026, reads: "We don't show you personalized ads based on your content from Drive, Gmail, or Photos."1

The after is the same line — with a door in it. The version marked "Effective October 16, 2026" reads: "We don't show you personalized ads based on your content from Drive, Gmail, or Google Photos, unless you choose to make data from these services available to other Google services that support personalized ads."2

Google publishes its own comparison of the two versions. It marks 8 insertions and 5 deletions, and the "unless" clause is one of the insertions.3 A Hacker News thread posted October 23, "Now Google says it might use your Gmail and Google Photos data to target ads," had drawn 378 points and 195 comments by October 25.4

The clause matters to anyone who keeps notes, drafts or records in a connected service — and the moment of decision is the connect prompt.


What changed in Google's ads sentence on October 16, 2026?

Google's ads sentence gained an exception. Content from Drive, Gmail and Photos was flatly excluded from personalized ads through the October 1, 2026 version. From October 16 it is excluded "unless you choose to make data from these services available to other Google services that support personalized ads."2 Connecting the apps is the choice.

The comparison page shows a second, quieter edit in the next paragraph. The old text said Google may show personalized ads "based on your interests and activity across Google services."1 The new text reads "based on your interests, activity, or personalized content from Google services."2 Personalized content — not only activity — is now named as an input.

A new passage explains the path. "You may choose to allow certain Google services, including services like Search or Gemini Apps, to access data from other Google services you use, such as Google Photos and Workspace apps like Gmail and Drive," it says, adding: "For example, Search may process some of the data that you allow Search to access to personalize content, including ads."2

The glossary names the feature. "If you choose to allow certain services to process data from other services, including for features such as Personal Intelligence, the same data may be used to personalize your experience."2

Four versions of the policy have taken effect in 2026 so far: April 2, May 26, October 1 and October 16.5 The date at the top is the cheapest thing on the page to check.

Does Gemini use my Gmail for ads? Google gives two answers

Google has two answers, one per product. The Gemini Apps Privacy Hub says your Gemini Apps chats "are not being used to show you ads."6 The new ads path runs through Search services — AI Mode and Ask Maps — when you connect content apps through Personal Intelligence. There, summaries and inferences drawn from Gmail and Photos may shape ads.7

Start with the Search side — it is the one the October change describes. Google's help page for Personal Intelligence says: "Ads will not be based directly on your Gmail inbox or your Google Photos library."7

The next two sentences carry the weight. "If you connect your content apps to Search Services, Search may reference summaries, excerpts and inferences from these sources to personalize your experience and do things for you. That same info may be used to show you ads, so they are even more helpful for you."7

So the honest summary is neither "Google reads your email for ads" nor "nothing changed." The inbox is not the input. What Search derives from it — a summary, an excerpt, an inference — can be.

The same page adds that "advertisers will not have access to your Gmail inbox or Google Photos," and names the switch: "You can adjust your settings, including disconnecting content apps from Search services, at any time."7

The Gemini side reads differently. "Your Gemini Apps chats are not being used to show you ads. If this changes, we will clearly communicate it to you."6 That hub was last updated October 23, 2026 — a week after the policy change, so both statements are current at once.

Does Copilot use my prompts and files for ads?

Yes, and the useful detail is which files. Microsoft's privacy statement says Copilot "uses prompts and related data to provide and improve services, including relevant advertising."8 Files stored on your device or in cloud storage are excluded from ad targeting. Files you share into a Copilot conversation are named as ad-relevant.8

The advertising section of the statement draws the line in two sentences. First the exclusion: "Microsoft does not use what you say in email, human-to-human chat, video calls or voice mail, or your documents, photos, or other personal files stored on your device or cloud storage to target ads to you."8

Then the inclusion: "Ads may be shown that relate to the current interaction you are having with us, such as your Copilot conversations (including files you share); your current location; transactions; product usage; search queries; or the content you're viewing."8

For a note-keeper this is the most practical line in the whole audit. The trigger is not storage — it is the act of attaching a document to a chat.

Microsoft's support page splits ads into two kinds. "Generic ads are selected based on the most recent content of your current conversation and don't use your past chat history. Personalized ads might use your chat history, saved memories, and other Microsoft data you've chosen to share to tailor ads and offers to your interests over time."9

The switch it names is Settings, then Personalization, then the "Allow ads personalization" toggle. The page also says who sees what: "If you use Microsoft Copilot and don't have a Microsoft 365 subscription, you might see advertising in Copilot," and "Regardless of your settings, Copilot doesn't show personalized advertising to users under the age of 18."9

Does ChatGPT use my chats for ads?

On the Free and Go plans, yes. OpenAI's help article says ads use the context of your current chat, and with personalization on, also other chats, ads history and, if memory is on, memories. Plus, Pro, Business, Enterprise and Edu accounts have no ads.10 The article says nothing about connected apps or uploaded files.

OpenAI announced the plan on January 16, 2026, in a post by Fidji Simo: "we plan to test ads at the bottom of answers in ChatGPT when there's a relevant sponsored product or service based on your current conversation."11 The same post says: "Plus, Pro, Business, and Enterprise subscriptions will not include ads."11

The help article dates the start: "Ad testing started in the United States on February 9, 2026."10 The "Testing ads in ChatGPT" post, updated August 11, 2026, describes the match: "During the test, we decide which ad to show by matching ads submitted by advertisers with the topic of your conversation, your past chats, and past interactions with ads."12

The personalization switch narrows that — it does not empty it. "If you turn off ads personalization, you will still see ads based on the context of your current chat thread, but we won't use other chat threads, ads history or topics to inform the ads we show you."10 The help article adds: "If memory is on, ChatGPT may save and use memories and reference recent chats when selecting an ad."10

The controls sit in Settings, under Ad Controls. Two narrower facts are worth knowing — one about a mode, one about a region. "Temporary Chats will not show ads," and "Personalized ads are not initially available in the European Economic Area (EEA) or Switzerland."10 If you delete your ads data, OpenAI says it stops being used for ads and is "retained for up to 30 days before being removed from our servers."10 The same article names one more switch on the Free plan: an Ads-Free mode that "removes ads, but comes with lower usage limits and reduced feature access."10

The gap OpenAI's pages leave open

Google and Microsoft each name what happens to content you connect or attach. OpenAI's ads help article names the current chat, past chats, memory and ad interactions — and stops there. It does not mention connected apps or uploaded files, and neither do the January 16 or February 9 posts.111210

That silence is not a yes — and not a no. The fair reading is that the ads pages do not address it. If you plan to connect a drive or upload a folder, that is the question to put to the provider, in writing.

One sourcing note, in plain view. OpenAI's live pages refused automated requests on the read date, so the quotes above were matched against archived copies of openai.com and a rendered copy of the live help URL. Open the help article yourself before relying on it — it is undated, and OpenAI says availability "may continue to evolve as testing expands."

Does Meta AI use my chats for ads?

Yes, since December 16, 2025. Meta announced on October 1, 2025 that it would "use your interactions with AI at Meta to personalize the content and ads you see, including things like posts and reels."13 Sensitive topics such as health, religion and political views are excluded. The newsroom post names no switch for this one signal.

Meta's own scale figure explains why this row matters: "More than 1 billion people use Meta AI every month."13 Meta said it would start notifying people on October 7, 2025, several weeks before the change took effect.13

The exclusion is specific. "When people have conversations with Meta AI about topics such as their religious views, sexual orientation, political views, health, racial or ethnic origin, philosophical beliefs, or trade union membership, as always, we don't use those topics to show them ads."13

There is one account-level carve-out: "if you have not added your WhatsApp account to an Accounts Center, interactions with Meta AI on WhatsApp will not be used to personalize experiences across different accounts on Meta Company Products."13 The controls Meta points to are Ads Preferences and Accounts Center linkage, not a toggle for AI interactions alone.

Region matters here too — the post says: "We're rolling out these changes in most regions."13 Most is not all.

Which AI assistants say they show no ads?

Four answers, none of them a ranking. Anthropic says Claude "will remain ad-free," in a February 4, 2026 post with a revisit clause. ChatGPT's Plus, Pro, Business, Enterprise and Edu plans have no ads. Copilot's ads apply to users without Microsoft 365. Gemini Apps chats are not used for ads. Each answer has its own scope.141096

Anthropic's statement is the most direct of the group. "So we've made a choice: Claude will remain ad-free. Our users won't see "sponsored" links adjacent to their conversations with Claude; nor will Claude's responses be influenced by advertisers or include third-party product placements our users did not ask for."14

Read it at its actual weight. It is a public commitment in a blog post — not a term in a contract — and the same post adds: "Should we need to revisit this approach, we'll be transparent about our reasons for doing so."14

It also covers ads inside Claude — not Anthropic's own marketing. Anthropic's privacy policy, effective September 10, 2026, says: "You can opt-out of sharing your personal data for targeted advertising to promote our products and services, and we will honor global privacy controls."15 That is a different sentence about a different thing.

The other "no ads" answers carry scope words. ChatGPT's is tied to the plan you are on. Copilot's page ties ads to not having a Microsoft 365 subscription. Gemini's is tied to the Gemini Apps product, not to Search.

Side-by-side: the ads sentence, six rows

The table puts each provider's ads sentence in one row: what the page says can shape ads, the switch it names, who sees no ads, and the page's own date. Google takes two rows because Search and Gemini Apps have different answers. Rows are alphabetical. The table compares sentences, not products, and ranks nothing.

AssistantWhat the page says can shape adsThe switch the page namesWho sees no adsPage's own date
ChatGPT (Free, Go)Current chat; with personalization, other chats, ads history; memories if memory is on10Settings > Ad Controls; Temporary Chats show no ads10Plus, Pro, Business, Enterprise, Edu; under-18s; Free users who switch to Ads-Free, with lower limits10Help article undated; posts January 16 and February 9 (updated August 11), 2026
ClaudeNothing in Claude, by stated commitment14—All users, with a revisit clause14February 4, 2026
Copilot (consumer)Prompts and related data; current conversation "including files you share"; with personalization, chat history and saved memories89Settings > Personalization > "Allow ads personalization"9Ads named for users without Microsoft 365; no personalized ads under 189"Last Updated: September 2026"
Gemini AppsChats "are not being used to show you ads"6——"Last updated: October 23, 2026"
Google Search with Personal Intelligence"summaries, excerpts and inferences" from connected Gmail and Photos; not "directly" the inbox or library7Disconnect content apps from Search services7—Policy "Effective October 16, 2026"
Meta AI"your interactions with AI at Meta," from December 16, 202513Ads Preferences and Accounts Center; no signal-specific switch namedNo group named; sensitive topics excluded13Newsroom, October 1, 2025

Read across the second column and the pattern is plain. The trigger is different in every row: a connection, a prompt, an attached file, a past chat, a memory, an interaction — or nothing at all.

Why search for one sentence instead of reading the policy?

Because almost nobody reads the whole thing — and the research says so. In a study of 543 participants, Jonathan Obar and Anne Oeldorf-Hirsch found 74% skipped the privacy policy. The policy "should have taken 29–32 minutes" to read; those who did open it spent an average of 73 seconds.16

The study appeared in Information, Communication & Society in 2020. Its abstract reports that "74% skipped PP" and that, "for those that didn't select the clickwrap, average PP reading time was 73 seconds."16 The finding is not a moral failing — it is a reasonable response to a document's length.

The documents in this audit are long. By our count of the text on each page, including navigation, Google's privacy policy runs to about 9,000 words and Microsoft's privacy statement to over 30,000. The counts depend on the extraction method; the order of magnitude does not.

A search box does what a reader cannot. Two words — "ads" and "advertis" — land you on the sentence that matters for this decision in under a minute.

How to find the ads sentence in two minutes

Open the provider's privacy policy and its assistant help page. Note the date at the top. Search the page for "ads" and "advertis." Find the sentence that names your input — a connected app, a prompt, a file you share, a memory. Then find the switch it names, and connect the narrowest scope that does the job.

The steps, in order:

  1. Note the date. Write down the effective or last-updated date. Google's policy changed four times in 2026; the date is how you know whether your last reading still holds.5
  2. Search, don't scroll. Use find-in-page for "ads" and then "advertis." The second catches "advertising" and "advertisers."
  3. Find the noun that matches what you are about to hand over. Connected apps (Google Search). Files you share into a chat (Copilot). Past chats and memory (ChatGPT, Copilot). AI interactions (Meta).
  4. Find the switch. A toggle, a disconnect option, a plan boundary. If the page names none, note that too.
  5. Connect the narrowest scope. One folder, not the drive. The case for scoping access is laid out in AI Agents Can Now Reach Into Your Files.

The ads sentence is not the only one worth finding. The other is what happens to your data if the company is sold — see Your AI Chats Can Be Sold With the Company, which audits the business-transfers clause across five providers.

Re-check when the date changes — a clause read once is a snapshot.

Who should read the ads sentence first — and who can skip it

Read it first if you are about to connect a mail account, a drive or a notes folder, or if you attach documents to chats. Read it if you use a free or ad-supported plan. You can skip it if you use a plan or product its page puts outside ads — and you have checked the date.

Read the sentence before you connect if:

You can reasonably skip it if:

The honest caveat: a "no ads" line answers one question. Training, human review and business transfers are separate sentences in the same policies.

Frequently Asked Questions

Does Gemini use my Gmail for ads?

Not through Gemini Apps chats. Google's Gemini Apps Privacy Hub, last updated October 23, 2026, says those chats "are not being used to show you ads."6 The ads path is Search: if you connect Gmail and Photos to Search services through Personal Intelligence, "summaries, excerpts and inferences" from them may shape ads — though not "directly" the inbox.7

Does Copilot use my prompts for ads?

Yes. Microsoft's privacy statement says Copilot uses "prompts and related data to provide and improve services, including relevant advertising."8 Files stored on your device or cloud storage are excluded from ad targeting, but your Copilot conversations "(including files you share)" can relate to the ads shown.8 The switch is Settings, Personalization, "Allow ads personalization."9

Does ChatGPT use my chats for ads?

On the Free and Go plans, yes. With personalization off, ads still use "the context of your current chat thread," but not other chats or ads history.10 With memory on, memories and recent chats may be used. Plus, Pro, Business, Enterprise and Edu accounts have no ads, and Temporary Chats show none.10

Does Meta AI use my chats for ads?

Yes, since December 16, 2025. Meta said it would "use your interactions with AI at Meta to personalize the content and ads you see."13 Conversations about health, religion, sexual orientation, political views and similar topics are excluded from ads. The newsroom post names no switch for AI interactions alone, and the rollout covers "most regions."13

Which AI assistants have no ads?

Claude, by Anthropic's stated commitment that it "will remain ad-free," with a clause allowing a revisit.14 ChatGPT's Plus, Pro, Business, Enterprise and Edu plans.10 Microsoft says users without a Microsoft 365 subscription "might see advertising in Copilot."9 Gemini Apps chats are not used for ads.6 Each answer has its own scope — none is a ranking.

How do I turn off ad personalization in ChatGPT or Copilot?

In ChatGPT, open Settings, then Ad Controls, and turn off personalization; ads based on the current chat continue.10 In Copilot, open Settings, select Personalization, and turn off "Allow ads personalization"; generic ads based on the current conversation continue.9 Neither personalization switch removes ads; ChatGPT's Free plan also offers an Ads-Free mode, with lower usage limits.10

Did Google start using Gmail and Google Photos data to target ads?

Google's October 16, 2026 policy added an exception to its Drive, Gmail and Photos ads exclusion, covering data you make available to services that support personalized ads.2 Its help page says ads "will not be based directly on your Gmail inbox or your Google Photos library," but connected-app summaries and inferences may be used.7


The ads sentence is short, it is dated, and it is the one line on the page written for the moment you click connect.


MNMNOTE keeps your notes as Markdown on your own device; if you use its bring-your-own-key AI, the text you send goes to the provider you chose, under that provider's policy — which is exactly the sentence above.

Footnotes

  1. Google, "Privacy Policy," archived version effective October 1, 2026; the same sentence appears in the versions effective May 26, 2026 and July 1, 2025. https://policies.google.com/privacy/archive/20261001?hl=en-US (accessed October 25, 2026) ↩ ↩2

  2. Google, "Privacy Policy," effective October 16, 2026. https://policies.google.com/privacy?hl=en-US (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6

  3. Google, "Privacy Policy" version comparison, October 1, 2026 to October 16, 2026. https://policies.google.com/privacy/archive/20261001-20261016?hl=en-US (accessed October 25, 2026) ↩

  4. Hacker News, "Now Google says it might use your Gmail and Google Photos data to target ads," posted October 23, 2026; 378 points and 195 comments on October 25, 2026. https://news.ycombinator.com/item?id=50214218 ↩

  5. Google, "Privacy Policy," archived versions list. https://policies.google.com/privacy/archive?hl=en-US (accessed October 25, 2026) ↩ ↩2

  6. Google, "Gemini Apps Privacy Hub," last updated October 23, 2026. https://support.google.com/gemini/answer/13594961?hl=en (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6

  7. Google Search Help, "Personal Intelligence." https://support.google.com/websearch/answer/16859283?hl=en (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8

  8. Microsoft, "Microsoft Privacy Statement," last updated September 2026. https://www.microsoft.com/en-us/privacy/privacystatement (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7

  9. Microsoft Support, "Microsoft Copilot for individuals: your privacy controls and choices." https://support.microsoft.com/en-us/privacy/microsoft-copilot/privacy-controls (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9

  10. OpenAI Help Center, "Ads in ChatGPT." https://help.openai.com/en/articles/20001047 (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17

  11. Fidji Simo, OpenAI, "Our approach to advertising and expanding access to ChatGPT," January 16, 2026. https://openai.com/index/our-approach-to-advertising-and-expanding-access/ (accessed October 25, 2026) ↩ ↩2 ↩3

  12. OpenAI, "Testing ads in ChatGPT," originally published February 9, 2026, updated August 11, 2026. https://openai.com/index/testing-ads-in-chatgpt/ (accessed October 25, 2026) ↩ ↩2

  13. Meta, "Improving Your Recommendations on Our Apps With AI at Meta," October 1, 2025. https://about.fb.com/news/2025/10/improving-your-recommendations-apps-ai-meta/ (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10

  14. Anthropic, "Claude is a space to think," February 4, 2026. https://www.anthropic.com/news/claude-is-a-space-to-think (accessed October 25, 2026) ↩ ↩2 ↩3 ↩4 ↩5 ↩6

  15. Anthropic, "Privacy Policy," effective September 10, 2026. https://www.anthropic.com/legal/privacy (accessed October 25, 2026) ↩

  16. Jonathan A. Obar and Anne Oeldorf-Hirsch, "The biggest lie on the Internet: ignoring the privacy policies and terms of service policies of social networking services," Information, Communication & Society 23(1): 128–147, 2020. https://doi.org/10.1080/1369118X.2018.1486870 ↩ ↩2