Is My Phone Listening to Me? Why Do I See Ads for Things I Talk About?

Example scenario: over dinner, you and a friend talk about camping tents. Then, the next morning, you open Instagram and the first thing you see is a tent ad. You wonder, “Is my phone listening to me?” You are not the only one asking. As the team that builds ad campaigns, we also hear this question all the time.

In this guide, we work through the question like detectives, one clue at a time. Talha Aslan and team build campaigns in Meta and Google ad platforms every day, so we know which signals sit behind an ad. First, we look at the evidence. Then we walk through the real causes, a 7-day test you can run yourself and the settings that control your ad profile.

What do the platforms officially say about the microphone?

The first clue comes from the platforms themselves. Meta published an official statement in its newsroom on exactly this question. In it, Meta says Facebook does not use your phone's microphone to inform ads or to change what appears in your feed. Specifically, the app only accesses the microphone with your permission. Moreover, you have to be using a feature that needs audio, such as recording a video.

Google makes a similar point. Its Safety Center page on ads and data asks whether Google listens to your phone calls to show you ads. The answer is a plain no, because Google says it never shows ads based on what you say over the phone. The same page also lists the data it does use: sites you visit, apps you use, your searches and your location. In other words, your conversations are not on that list.

Voice assistants, on the other hand, are a separate worry. Google's Assistant help page says the assistant waits in standby mode until it detects an activation, such as “Hey Google”. In standby, Google says, what you say does not go to its servers. However, Google also admits that a similar sound can trigger the assistant by mistake. So it is worth reviewing your assistant settings now and then.

What did independent research find?

Company statements are not enough on their own, so independent checks matter. For example, researchers at Northeastern University ran a study called Panoptispy. It examined how more than 17,000 Android apps handle images, video and audio. The researchers write that in their tests, no apps appeared to send out audio.

However, the same study uncovered a different risk. Some apps, and third-party code bundled inside them, sent screenshots and screen recordings to other parties without telling the user. So we are not saying privacy risks do not exist; they often sit somewhere other than where people look.

The project page also adds another detail. In the past, researchers caught some apps passively listening for ultrasonic audio beacons that humans cannot hear. That method targeted inaudible signals, not conversations. Still, it shows the microphone can serve purposes beyond speech, which is why we take permission lists seriously.

Every study also has limits. This one focused on Android apps under specific test conditions. That is why we do not claim that no app ever listens. Instead, we can say that the evidence does not support the listening theory.

Why is secret listening a poor deal for advertisers?

Treat this as an advertiser's cost estimate rather than technical proof. First, recording audio nonstop, transcribing it and making sense of it would take serious processing power. On top of that, current iOS and Android versions show an indicator whenever the microphone is in use. Apps also have to ask for your permission before they touch the microphone.

Above all, casual conversation is a weak signal for an advertiser. For example, someone can talk about tents without any plan to go camping. By contrast, the following actions are cheap to collect and far more accurate.

• Typing “two-person tent” into a search box.
• Opening a tent product page and checking the price.
• Watching camping videos all the way to the end.
• Walking into an outdoor gear store with location services on.

Ad systems already collect plenty of signals like these. As a result, they can make surprisingly accurate guesses without listening at all. In our view, the risk of secret listening would far outweigh the small extra value it could add.

Do your messages and emails turn into ads?

This is the close cousin of the listening question. On the same Safety Center page, Google also says it doesn't use data from Gmail, Drive or Google Photos for ads, unless you choose to make that data available to other Google services that support ads. The exception is when you choose to make that data available to other Google services that support ads. In other words, that part depends on a choice you make.

For messaging apps, the only reliable source is the app's own official statement. So look at the app's privacy page and check whether it says message content feeds into ads. Trust the written policy over rumors. In practice, the ad platforms we use offer no option to target people based on message content.

Is my phone listening to me, or did I search first?

The first question in any investigation is timing. We often run a quick search shortly before or after a conversation, then forget it within minutes. For example, you might have tapped a link a friend sent you the day before. Or, mid-conversation, you might have checked the product's price on your phone.

The reverse also happens. Sometimes the conversation comes after the ad, not before it. An ad you scrolled past without noticing leaves a trace, and later you bring up the topic yourself. Then the same ad appears again the next day, and you flip the order in your head.

So the first check is simple: look at your browser history, your YouTube watch history and your in-app searches. Google says personalized ads draw on activity saved to your Google Account. That activity can come from any device where you sign in to your account. In other words, a search on your tablet can shape the ads on your phone.

How can someone else's search become your ad?

Example scenario: your partner researches a stroller model one evening. Meanwhile, you only join the conversation and search for nothing. Then, the next day, the same brand shows up on your phone. There are several possible routes here, and none of them needs a microphone.

• Shared accounts: if the living room TV runs the same YouTube account, what plays there can also shape your ad profile.
• The same home internet: devices on the same router can look similar from the outside.
• The same location: people in the same home or neighborhood may fall into the same local campaigns.
• Social ties: the accounts you follow and shared interests can pull your profile closer to people around you.

Platforms do not explain in detail how they weigh these signals. We do not see them at an individual level in the ad dashboard either. So we cannot say which one tips the balance; read them as possible explanations, not proven mechanisms. Still, every one of them is simpler than listening to a conversation. Finally, separate user profiles on shared devices are the easiest way to reduce the mix-up.

How does location history affect your ads?

Location is one of the strongest signals in advertising. For instance, Google says personalized ads can reflect your general area while you use its services. Even with personalization off, contextual signals such as your current location and the time of day also play a role.

Example scenario: on Saturday you browse a furniture store but buy nothing. That evening you and your partner talk about a new couch. Then, on Sunday, furniture ads start appearing. The stronger explanation is your visit near that store, plus local campaigns aimed at that area.

You can see which apps access your location in your phone's privacy settings. Turning off unnecessary location permissions reduces these matches, but it will not end them. That is because some ads only consider the city or neighborhood you are in.

How do the websites you visit remember you?

When you open a website, a small measurement code often runs in the background. Advertisers call it a pixel. Specifically, a pixel tells the ad platform that you viewed a page or added a product to your cart. Cookies, in turn, are small records your browser keeps so a site can recognize you on your next visit. We also explain how this works in our guide to the Meta Pixel.

Advertisers call ads built on this data retargeting. Example scenario: you look at a coffee machine, then leave without adding it to your cart. That evening, you talk about coffee machines at home. The next day, Instagram shows you the exact machine. The order feels like “conversation, then ad”, but the real order was “visit, then ad”. Our article on remarketing also shows how brands use this method.

Your answer in a cookie consent banner affects this flow too. For example, when you reject advertising and analytics cookies, fewer measurement signals reach the platforms. We cover how that choice affects advertising in our post on cookie consent and Google Consent Mode. Clearing cookies or blocking third-party cookies also reduces this kind of tracking. However, activity saved in the accounts you use stays where it is.

How do brands match your phone number and email to ads?

Shopping at a store or signing up for an app often means handing over your phone number or email. After that, the brand can upload this customer list to an ad platform, within legal limits. The platform then shows ads to the accounts that match the list. Meta calls this a Custom Audience, while Google calls it Customer Match.

Does that mean the brand shares your details in plain text? According to Google's help page, the brand can hash the email and phone data itself with an algorithm called SHA256. Hashing is a one-way method that turns data into a scrambled string. Otherwise, Google hashes the data with the same algorithm. The platform then compares these strings with the hashed versions of its own account data. Finally, once matching and policy checks finish, Google marks the file for deletion. Meta also has a dedicated help page about hashing customer information.

So seeing an ad from a brand you gave your number to is not surprising. It has nothing to do with your conversation and everything to do with your earlier relationship with that brand.

Why do you see ads for things you never searched for?

This is the most confusing case: “But I never searched for that!” Here, lookalike audiences come into play. Put simply, a lookalike audience is a targeting type that finds new people who behave like a brand's existing customers. The platform identifies what those customers have in common and then looks for similar profiles.

Example scenario: over recent weeks, you watched camping videos, followed nature photographers and liked an outdoor shop's post. You never searched for a tent, yet your behavior resembles people who buy tents. So when a tent ad appears on the day you mention camping, you link it to the conversation. Meanwhile, the system has been reading small signals for weeks.

Alongside lookalikes, interest estimates are also at work. For instance, the platform infers interests from the content you like and linger on. These guesses are sometimes spot-on and sometimes odd, but they all come from your behavior, not your voice.

What do apps learn about you from each other?

Many apps on your phone fill their ad space through ad networks. These networks can use an advertising ID to connect your activity across different apps. That ID is a unique identifier for advertising, and you can reset or delete it. That is exactly how Google describes it for Android.

Example scenario: you look at a bike helmet in a shopping app. A few hours later, a helmet ad appears in a game. Nothing in between listened to you; instead, an ad infrastructure linked the two apps. We also explain that infrastructure in plain language in our guide to programmatic advertising.

Apple gives users a clear choice here. According to Apple Support, apps must ask your permission before they track you across other companies' apps and websites. Android, on the other hand, lets you reset or delete the advertising ID. It can also note topics of interest based on the apps you use. We show where these settings live further below.

Why does the Baader-Meinhof phenomenon feel like eavesdropping?

The frequency illusion, also called the Baader-Meinhof phenomenon, makes something you just noticed seem to be everywhere. For example, after you talk about a new car model, you start spotting it in traffic constantly. The number of those cars did not change; your attention did.

Ads work the same way. You ignore most of the ads you see in a day, a habit advertisers call banner blindness. However, the one ad that matches your conversation jumps out at once. You forget the countless ads that matched nothing, and you retell the one that did for months.

The brain works like a pattern-finding machine. When it spots a coincidence, it looks for a cause. Then the explanation that sticks most easily is that someone must be listening. So the feeling is very real, even though it is not proof on its own.

Is coincidence more common than you think?

Let's run a rough example calculation. Say you see a few hundred ads in a day across social media, news sites and apps. Meanwhile, you talk about dozens of different topics that same day, at home, at work and on the phone. These numbers vary widely from person to person, so we use them only to show the logic.

With so many ads and topics, an occasional “that is exactly what we talked about” moment is not surprising. What would be surprising is never having one. Moreover, ads are not random. The system picks them based on your age, location and interests, which raises the odds of a match.

Seasons and news cycles add to this. For instance, when the weather heats up, everyone talks about air conditioners. Air conditioner brands also step up their ads in those same weeks. Back-to-school season brings backpacks into conversations, and the holidays bring gifts. Because conversation and advertising follow the same calendar, they collide often.

Which beliefs are myths and which are facts?

We put the beliefs we hear most often next to what we actually know. In short, the table sums up this guide at a glance.

Is my phone listening to me right now, and how can I tell?

Fortunately, phones do not hide microphone use. Apple's support page on the orange and green indicators in the iPhone status bar spells it out. Specifically, on iOS 14 or later, an orange dot appears when an app uses the microphone. On the other hand, a green dot means an app uses the camera, or the camera and microphone together. If the Differentiate Without Color setting is on, the orange indicator shows as a square instead.

Apple's iPhone User Guide also adds one more clue. If an app has recently used the camera or microphone, a message appears at the top of Control Center. So whenever you feel suspicious, open Control Center and check that message.

On Android, Google's help page on checking whether your camera or microphone is on describes a green indicator. When apps use your camera or mic, it appears in the top right corner of the screen. Then swipe down and tap it to see which app is using it. That said, some of these steps work only on Android 12 and up, and the look can vary by manufacturer.

Which settings let you turn off microphone access?

If doubts remain, the most solid step is to leave microphone access only with apps that truly need it. Menu names can change by version and manufacturer, so check the current steps on the official help pages.

1. On an iPhone, go to Settings > Privacy & Security > Microphone.
2. Then switch off every app in the list that does not need access.
3. On Android, open Settings > Security & privacy > Privacy > Privacy controls and find Microphone access.
4. For an app-by-app view, open the privacy dashboard to see which apps used which permission and when.
5. You can also use the microphone tile in Quick Settings to cut or restore access with one tap.

When you remove an app's microphone access, features such as voice messages or video recording may stop working. That is normal; you can also grant access again whenever you need it.

Is my phone listening to me? How can I test it in 7 days?

If you want to see for yourself, you can set up a simple experiment. It is not a scientific study; still, it is a good way to get to know your ad profile. Above all, keep the conversation as the only variable. Here are the steps of the 7-day experiment.

1. First, pick a product nobody around you owns and you have never searched for, such as a pottery wheel.
2. Do not search for it, type it or mention it in messages on any device. Also ask your household not to search for it.
3. Next, pick a second product as a control and search for it once in your browser.
4. Then, with your phone nearby, talk about the first product out loud a few times a day.
5. Log every related ad you see with the app name, the time and a screenshot.
6. For each ad, open the option that explains why you see it and note the reason.
7. Finally, on day seven, compare both products and review your microphone permissions.

To strengthen the test, run it with a friend. One of you only talks about the product, and the other only searches for it. Living in different homes makes the result clearer, because it removes shared network and shared account effects. Then, at the end of the week, compare both ad feeds; the difference shows which signal works harder.

How should you read the results?

One possible outcome: ads for the control product show up, while the spoken product never appears in ads. In other words, the pattern suggests ad systems react to behavior rather than conversation. If even the control product never appears, few brands may advertise it, so repeat the test with a different product.

What if an ad for the spoken product does appear? Then, before you panic, ask yourself these questions.

• Did anyone at home, or anyone using the same account, search for this product?
• Is the product seasonal or in the news right now?
• Does the ad explanation mention a broad reason such as age, location or interests?
• Could you have seen the same ad earlier without noticing it?

A one-week experiment by one person cannot deliver final proof. Still, it clearly shows which signals are strong in your profile. In practice, checking the microphone indicator and your permission list gives you the clearest answer.

Where can you see and change your ad profile?

The quickest way to understand why an ad reached you is through the platforms' own transparency tools. Menu names change over time, so check the items below against the current names on the official help pages.

• Google: My Ad Center shows what information Google uses for ads on Search, YouTube and Discover. There you can also adjust topic and brand preferences or turn off personalized ads.
• Meta: the Facebook Help Center has a page that explains why you see ads from a particular advertiser. In the ad's own menu, an option along the lines of “Why am I seeing this ad?” shows the reason. You can also manage interests and ad settings in the ad preferences section of your account settings.
• Apple: under Settings > Privacy & Security > Tracking, you decide whether apps can ask to track you. Turn off Allow Apps to Request to Track, and Apple says every tracking request then counts as declined.
• Android: in the privacy section of Settings, under Ads, you can reset or delete your advertising ID. On newer versions, you also reach the ad privacy settings through the Google section.

That said, according to Google, you will still see ads after turning off personalization; they simply become less relevant. If you wonder which ads a brand is running right now, you can also look at Meta's public archive. Our guide to the Meta Ads Library explains how to read it. Additionally, our ad library search tool scans several platforms in one search.

What does the ad dashboard show us, and what does it hide?

This section comes straight from our day-to-day work. When we set up a campaign, we describe an audience by region, age range and interests. The platform then shows us the estimated size of that audience, usually as a range. As the ad runs, we also see aggregate numbers like impressions, clicks and conversions.

Likewise, when we upload a customer list, we only get an aggregate figure for how many people matched. However, what we cannot see matters more to you.

• The names, messages or chats of the people who saw the ad.
• Which individual people on the list matched.
• Anyone's conversations; the dashboards we use have no option to target by conversation.

In short, we manage audiences as groups. We have no way to learn what an individual said, and no menu lets us request it. Reports also show the combined behavior of groups, not individuals. Even with location targeting, we choose an area; we never see who is where.

What were the “active listening” reports about?

Every so often, a headline on this topic makes waves. For example, one media company's sales pitch claimed it could target ads using what people said near device microphones. 404 Media published the pitch deck. According to the outlet, Google then removed the company from its advertising Partners Program.

We take two lessons from this story. First, a claim in a sales deck does not prove that the technology works or that it is widespread. Second, the claim conflicts with the major platforms' official statements, and Google's reaction fits that stance. We could not verify the deck's contents independently, so we call it neither a total fabrication nor proven.

So when you come across a similar story, ask three questions. Who is making the claim, and what are they selling? Is there an independent test or technical review? What do the named platforms officially say? In practice, the answers show how seriously to take the story.

A note for businesses: how can you use these signals transparently?

If you run a business, you can use the same signals without unsettling your customers. Here are a few principles we also apply in our own work.

• Collect cookie consent through a banner that offers a real choice, and respect visitors who decline.
• Explain your use of pixels and customer lists in plain language in your privacy notice.
• Avoid creepy personalization in ad copy, such as “the product you talked about yesterday”.
• Do not show the same ad to the same person endlessly; instead, set a sensible frequency cap.

Our guide to building a GDPR-compliant website covers what to watch for on your site. This is not legal advice; consult a lawyer for a binding assessment. Transparency is also a matter of trust. A creepy ad hurts the brand, and lost trust is hard to win back. So if you want campaigns built on these principles, take a look at our Meta ads management service.

So what should you do?

Seeing an ad for something you just talked about feels unsettling. However, behind it there is usually no microphone, only digital traces left by you and the people around you. Your searches, site visits, location, the phone number you gave a brand and lookalike behavior build that picture. Then the frequency illusion and plain coincidence add the finishing touch.

Your next steps are clear: learn the microphone indicator, trim your permissions and check your ad profile. So next time, ask yourself one question first: is my phone listening to me, or did I leave a trail?

Core guide: What Is Digital Marketing? Its types, examples and how it is done, on one page with an interactive film that travels through the brain.