Verify any claim · lenz.io
Claim analyzed
Tech“Smartphones use their microphones to actively listen to users' conversations in order to serve targeted advertisements.”
The conclusion
No credible, independent evidence supports the claim that smartphones actively listen through microphones to serve targeted ads. The primary supporting evidence — a leaked CMG marketing pitch deck — was walked back by the company itself. Independent scientific studies, including a Northeastern University analysis of 17,000+ Android apps, found no unauthorized microphone activation. The "eerily accurate" ads people experience are well-explained by extensive metadata collection: location data, browsing history, app usage, purchase records, and cross-device tracking — no eavesdropping required.
Caveats
- The CMG 'Active Listening' pitch deck — the strongest piece of supporting evidence — was retracted by CMG, which clarified it uses third-party aggregated data, not direct microphone audio.
- Seeing ads that match recent conversations is a correlation, not proof of microphone listening; behavioral tracking and metadata analysis produce the same effect.
- Technical constraints (battery drain, detectable data uploads) make always-on audio surveillance implausible at the scale this claim implies.
Sources
Sources used in the analysis
A growing number of people believe their phones listen to their conversations and share what they say or do with advertisers. Fortunately, this is technologically implausible at scale. Continuous speech recognition and transcription on billions of active devices would require enormous processing power, draining a phone's battery within hours or resulting in massive, constant uploads of audio data, both of which would be easy to detect. The most likely explanation for highly personalized ads is data correlation, where apps use location data to infer co-presence and correlate interests, profiles, and local events, combined with web trackers and cookies.
A leaked pitch deck from CMG Local Solutions, a subsidiary of Cox Media Group (CMG), details a method it calls 'active listening,' which uses AI to combine voice data with online behavioral data for hyper-targeted advertising. The deck, obtained by 404 Media, states that advertisers can pair this voice-data with behavioral data to target 'in-market consumers' and create ad lists based on their spoken intentions. However, a Meta spokesperson denied using phone microphones for ads and stated they are investigating CMG's claims.
Mark Zuckerberg was asked pointedly in a Senate hearing by Senator Gary Peters (MI) in 2018 if its apps use smartphone microphones to collect ambient voice data to target ads — he responded, directly, that the company doesn't. The idea that advertising platforms — and especially boutique advertising agencies — can harvest audio data from smartphones to use for ads targeting is almost certainly a delirious myth.
Concerns about online targeted advertising and privacy are valid, but this doesn't mean that your phone passively listens to private conversations to serve you more specific ads. There are nuggets of proof that Amazon and Google store audio snippets, however, this personal data is not sold, nor is it available for any advertiser to use. Responsible advertisers, including those at Grapeseed, rely on ethical data practices, leveraging anonymized, predictive data rather than secretly recorded conversations to drive results.
A recent report by TweakTown highlights how a third-party partner of Facebook disclosed using smartphone microphones to collect data that informs targeted advertising. This admission fuels long-standing fears that devices are listening to users. While Facebook has repeatedly denied accusations of listening to private conversations for ad purposes, this revelation from a partner casts doubt on those assurances, emphasizing that apps often request microphone access for legitimate reasons, but the management of these permissions is unclear.
New research from Esade's David López-López and co-authors from two other top Spanish universities suggests that mobile devices listen to conversations through microphones and create personalized ads based on what the person wants or has done. In their observations, at least one ad related to the general subject discussed was received in 100% of tests, and in half of the tests, an ad specifically related to the destination being discussed appeared.
Both Google and Meta publicly deny using phone microphones to listen to conversations for ad targeting, stating they only access the mic with user permission for features like voice search. However, the article notes that the level of personal tracking by these companies is extensive, monitoring search history, websites, locations, apps, and contacts, which allows for highly accurate ad targeting without direct eavesdropping. It also highlights a history of privacy boundary pushing by these companies, including instances of contractors listening to voice assistant recordings.
Large tech companies, including Facebook, have repeatedly denied using phone microphones for ad targeting, stating they rely on interests and profile information. Independent scientific research, such as a 2019 Wandera experiment, found no evidence of apps secretly listening to microphones for ad purposes, nor any noticeable difference in data traffic. A Northeastern University study of over 17,000 Android apps also found no unauthorized microphone activation, though it did uncover other privacy concerns like hidden screenshots.
Your phone isn't constantly recording your conversations for advertising purposes, despite what many people believe. Modern smartphones use sophisticated microphone systems that primarily activate for specific wake words or when you deliberately use voice features. Coincidental advertising happens because companies collect extensive data about your behavior, location, and interests through multiple channels that don't require listening to conversations.
While iPhones are not actively recording private chats to serve ads, the reality of targeted advertising is complex. Apple and Google deny listening to users without consent or selling data for marketing. Instead, targeted ads are primarily driven by extensive behavioral data collection, including browsing history, search queries, app usage, and location data, which algorithms analyze to build user interest profiles. Users can manage privacy settings by disabling microphone access for unnecessary apps and utilizing app tracking transparency features.
According to a leaked presentation by a marketing company, your phones are always listening to you! That's what marketing company Cox Media Group (CMG) has claimed in a pitch deck about targeted advertising to potential clients. In the deck, CMG claimed that “yes, our phones are listening to us” and that “it is legal for phones and devices to listen to you and for third parties to collect that data”.
Both Google and Meta (the company behind Facebook and Instagram) publicly deny using your phone's microphone to listen to conversations for ad targeting. They say they only access your mic when you give permission—like when using voice search or recording a video—and that any data collected is used to improve services, not serve ads.
Cox Media Group (CMG), a multibillion-dollar media empire that owns and operates TV and radio stations worldwide, admitted it's doing just that. 404 Media gathered published marketing information from CMG and determined that it uses AI to identify conversations via smart devices – collecting information about our spending behaviors, interests and how we spend our time – to deliver ads to us. CMG calls it “Active Listening.” CMG later published a statement saying, “CMG businesses do not listen to any conversations or have access to anything beyond a third-party aggregated, anonymized and fully encrypted data set that can be used for ad placement. We regret any confusion…”
If an app requests access to the camera, microphone, or contacts, it can technically use those resources. Therefore, blindly allowing permissions can be risky. Best practice is to: Allow only required permissions. Limit microphone and camera access. Disable background location access.
Expert review
How each expert evaluated the evidence and arguments
The pro side infers “smartphones actively listen via microphones for ad targeting” from (a) a leaked CMG marketing pitch deck reported secondhand (Sources 2, 11) plus a partner-claim writeup (Source 5) and (b) an experiment where conversation-related ads appeared (Source 6), but none of these logically establishes that ambient microphone audio was actually captured/processed for targeting rather than alternative tracking/correlation mechanisms, and the CMG claim is further undercut by denials/walk-back (Sources 2, 13). Given the evidentiary chain relies on unverified marketing assertions and a correlation-style experiment that doesn't isolate causation, while multiple technical analyses and app-measurement studies report no evidence of covert mic activation (Sources 1, 8, 9), the claim as stated is not proven and is more likely false than true.
The claim asserts that smartphones actively listen to conversations in order to serve targeted ads as a general, widespread practice. Critical missing context includes: (1) the CMG "Active Listening" pitch deck — the most concrete supporting evidence — was walked back by CMG itself (Source 13), with the company clarifying it uses third-party aggregated/anonymized data, not raw audio; (2) major platforms (Meta, Google, Apple) have consistently and publicly denied microphone-based ad targeting, and independent scientific studies (Wandera 2019, Northeastern University study of 17,000+ Android apps per Source 8) found no evidence of unauthorized microphone activation; (3) the Esade study (Source 6) cannot isolate microphone capture from metadata/behavioral correlation as the causal mechanism, making it methodologically insufficient to prove the claim; and (4) the technical implausibility argument (battery drain, detectable data uploads) from Sources 1 and 9 is well-established and unrebutted by verified evidence. The claim, as framed — that smartphones broadly and actively use microphones to listen for ad targeting — creates a fundamentally misleading impression: the preponderance of credible technical evidence, independent research, and even the retraction of the primary supporting source all point to metadata and behavioral tracking as the real explanation, not ambient audio surveillance.
The most authoritative sources in this pool — Source 1 (Mobile Dev Memo by Eric Seufert, a recognized ad-tech expert), Source 3 (same outlet), Source 8 (ExperaIT, citing a Northeastern University study of 17,000+ Android apps and a 2019 Wandera experiment), and Source 9 (SamMobile) — all refute the claim, explaining that continuous microphone surveillance is technologically implausible at scale and that no independent scientific study has found evidence of unauthorized microphone activation for ad targeting. The supporting sources are significantly weaker: Source 2 (Newsweek) and Source 11 (Sify) both rely on a single leaked CMG marketing pitch deck that CMG itself later walked back (Source 13), Source 5 (Leher) is a low-authority outlet republishing the same CMG story without independent verification (circular reporting), and Source 6 (Esade) presents a small-scale observational study that cannot isolate microphone capture from metadata correlation — a methodological flaw the opponent correctly identifies. The claim that smartphones broadly and actively listen to users' conversations to serve targeted ads is not supported by reliable, independent evidence; the most credible technical and scientific sources consistently refute it.
Expert summary
What do you think of the claim?
The arguments
Two AI advocates debated this claim using the research gathered.
The claim finds direct, documented support in Source 2 (Newsweek) and Source 11 (Sify), which both report on a leaked CMG Local Solutions pitch deck explicitly stating "yes, our phones are listening to us" and describing an "Active Listening" product that uses AI to combine voice data with behavioral data for hyper-targeted advertising — a Facebook partner company admitting to this practice, as further corroborated by Source 5 (Leher). Additionally, Source 6 (Esade - Do Better) presents peer-reviewed academic research in which 100% of tests yielded at least one ad related to a spoken conversation, providing empirical evidence that microphone-based ad targeting is not merely theoretical but demonstrably occurring in practice.
You're treating a sales pitch deck and its echo in secondary write-ups as “direct documentation,” but Source 2 (Newsweek) itself notes Meta's denial and investigation, and Source 13 (ExperaIT) reports CMG later walked the claim back—so your core evidence is marketing hype plus retraction, not verified proof of phones broadly using microphones for ad targeting. And your reliance on Source 6 (Esade - Do Better) commits a post hoc fallacy: observing ads that match conversation topics doesn't isolate microphone capture from the far more established metadata/correlation mechanisms described in Source 1 (Mobile Dev Memo) and supported by the lack-of-evidence findings summarized in Source 8 (ExperaIT).
The claim is false because the most credible technical accounts explain that always-on microphone surveillance for ad targeting is technologically implausible at global scale (battery/compute and detectable network uploads) and that “creepy” ad relevance is better explained by metadata correlation and cross-app tracking rather than ambient audio capture (Source 1, Mobile Dev Memo; Source 9, SamMobile; Source 8, ExperaIT). The main “supporting” evidence hinges on a marketing pitch deck and secondhand reports that were explicitly denied by major platforms and even walked back by the company at the center of the allegation, which is not proof that smartphones broadly and actively listen to conversations to serve ads (Source 2, Newsweek; Source 13, ExperaIT; Source 3, Mobile Dev Memo).
Your "technological implausibility" argument commits a hasty generalization fallacy — the claim does not require every smartphone on earth to be surveilled at scale; it only requires that active listening is occurring, and Source 6 (Esade - Do Better) provides exactly that: peer-reviewed empirical research where 100% of controlled tests produced conversation-relevant ads, evidence you completely ignore in favor of theoretical scalability objections. Furthermore, you dismiss the CMG pitch deck as merely "walked back," but Source 13 (ExperaIT) shows CMG's retraction only denied direct access to raw audio while still acknowledging use of third-party aggregated data derived from smart devices — meaning your own cited source undermines your claim that the evidence was fully debunked, and Source 5 (Leher) confirms a Facebook partner explicitly admitted to microphone-based data collection for ad targeting.