Table of Contents
Smart Home Devices Are Always Listening — What That Means for Kids
Alexa, Google Home, and similar devices record household conversations. Here's what parents need to know about what's captured, who sees it, and how to reduce exposure.
If you own an Alexa, Google Nest, or Apple HomePod, there is a microphone in your home that is always active. The device is engineered to listen for its wake word — “Alexa,” “Hey Google,” “Hey Siri” — and to activate recording when it hears one. This is how these devices work by design. What is less commonly understood is that the “always listening” state generates its own data trail: false triggers, recordings transmitted before the wake word is fully detected, and audio that ends up in cloud servers associated with your account.
For adults who’ve read the terms of service and made an informed choice, this is a calculated tradeoff. For children who live in households with these devices, it’s simply the environment they inhabit — and the question of what their voices, their conversations, and their behavioral patterns mean for their privacy is one that most parents haven’t fully examined.
Key Takeaways
- Smart speakers are technically “always on” and regularly capture audio that wasn’t intended as a command — accidental recordings are well-documented
- The Children’s Online Privacy Protection Act (COPPA) requires explicit parental consent before companies can collect voice data from children under 13, but enforcement has been inconsistent
- Amazon, Google, and Apple all retain voice recordings with varying default periods and allow user deletion, but data practices differ by platform
- FTC action in 2023 against Amazon included a $25 million penalty for COPPA violations related to Alexa’s retention of children’s voice data
- Practical steps — microphone muting, reviewing and deleting voice history, and using Kids Mode where available — meaningfully reduce exposure
How Smart Speakers Actually Work
The Wake Word Model
Smart speakers use a two-stage process. The first stage happens entirely on the device: a small neural network continuously analyzes audio for sounds resembling the wake word. This on-device processing theoretically doesn’t send data to the cloud. The second stage begins when the wake word is detected: the device begins transmitting audio to the company’s servers for processing.
The practical problem is imprecision. The on-device detector is trained on millions of speech samples but produces both false negatives (missing genuine commands) and false positives (activating on sounds that merely resemble the wake word). Every false positive is a brief audio clip of your household’s ambient conversation that gets transmitted to Amazon, Google, or Apple’s servers.
What Gets Recorded and Stored
Amazon, Google, and Apple have all faced scrutiny for using human reviewers to listen to voice recordings to improve accuracy. The practice was widespread before public reporting in 2019 prompted policy changes. Current default practices vary:
Amazon Alexa: By default, recordings are stored indefinitely but can be auto-deleted after 3 or 18 months via settings. Human review of recordings now requires opt-in. Amazon’s Help to Improve Alexa setting is what enables this review.
Google Assistant: Google stores audio activity associated with your account. Deletion can be automated via settings. Google has also faced scrutiny for contractor review of recordings.
Apple Siri: Apple’s approach has historically involved more on-device processing. After a 2019 whistleblower report about Siri recordings being reviewed by contractors, Apple revised its policies to require opt-in for human review.
| Platform | Default Storage | Auto-Delete Option | Human Review |
|---|---|---|---|
| Amazon Alexa | Indefinite | 3 or 18 months | Opt-in required |
| Google Assistant | Account activity | Customizable | Opt-in required |
| Apple Siri | More on-device | Account-based settings | Opt-in required |
| Amazon Echo Kids | Different retention rules | Yes | Restricted |
The COPPA Problem with Voice Assistants
The Children’s Online Privacy Protection Act prohibits operators of child-directed websites and online services from collecting personal information — including audio — from children under 13 without verifiable parental consent. Voice data is personal information under COPPA.
In 2023, the FTC and the Department of Justice reached a settlement with Amazon that included $25 million in civil penalties specifically for COPPA violations related to Alexa’s retention of children’s voice recordings. The FTC’s complaint alleged that Amazon retained children’s voice recordings longer than necessary, failed to honor parent deletion requests, and used the data to train its Alexa algorithm without required consent.
This enforcement action confirmed what privacy advocates had argued: that general-purpose voice assistants used in family homes were collecting children’s data without adequate consent mechanisms, and that the companies were aware of this.
Specific Concerns When Children Use Voice Assistants
Behavioral Profiling
Voice recordings of children provide rich data for behavioral profiling. Vocal analysis can infer mood, health status, and developmental stage. Over time, a corpus of a child’s voice provides a behavioral baseline more detailed than most parents realize. This data, associated with a household account, is retained in corporate databases with retention policies that have repeatedly shown gaps in enforcement.
Conversations Children Have Near Devices
False-trigger recordings don’t require the child to be talking to the device. A child’s conversation with a parent, arguments, emotional moments, and sensitive disclosures can all appear in household voice data. Amazon’s own leaked recordings from investigator reports have included domestic arguments, health conversations, and intimate moments that users had no idea were captured.
Kids Mode and Its Limitations
Amazon and Google both offer children-specific modes (Amazon Kids, Google’s family settings) with different content filters and, in Amazon’s case, different data retention policies. These are meaningful improvements over default settings. However, they require active setup, apply only when the child is specifically interacting with the device in the designated mode, and don’t prevent accidental capture of background conversations.
What Parents Can Do
Step 1: Review Your Voice History
Every platform offers access to your voice recording history. Doing this once is instructive:
- Amazon: Settings > Alexa Account > Alexa Privacy > Review Voice History
- Google: myactivity.google.com > Filter by “Voice & Audio”
- Apple: Settings > [your name] > iCloud > Siri & Search
Listen to several recordings. The number of false triggers, and the content captured, is often surprising.
Step 2: Delete Existing Recordings
From the same settings menus, you can delete individual recordings or all recordings associated with your account. Set up automatic deletion:
- Amazon: Settings > Alexa Privacy > Manage Your Alexa Data > Automatically delete recordings (choose 3 months)
- Google: myactivity.google.com > Auto-delete (choose 3 months)
Step 3: Opt Out of Human Review
Ensure you’re not opted in to allowing human contractors to review your audio:
- Amazon: Alexa Privacy > Manage Your Alexa Data > Help Improve Alexa (turn off)
- Google: Activity Controls > Voice & Audio Activity (manage settings)
Step 4: Use Physical Muting
All smart speakers have a physical mute button that disables the microphone hardware. This is the only method that provides certainty that the device isn’t listening. Consider making a practice of pressing mute when having conversations you wouldn’t want recorded — medical discussions, financial conversations, sensitive family discussions.
Step 5: Reposition or Remove Devices from Children’s Spaces
Bedrooms and children’s playrooms are rooms where conversations are private by nature. A smart speaker in a child’s bedroom captures everything said in that room. Consider whether the utility justifies the positioning.
Having the Conversation with Your Kids
Children old enough to use a voice assistant are old enough to understand, in basic terms, how it works. This is an excellent teaching moment for a broader conversation about digital privacy — what technology collects, who sees it, and how to make informed choices about what we share.
For younger children: “Alexa has a microphone that’s always on. When we talk to her, she listens and sends what we say to Amazon’s computers. That’s how she understands us.”
For older children: “Every time we use Alexa or Google Home, a recording goes to their servers. It usually stays there for a while. So it’s worth thinking about what we say near these devices the same way we’d think about anything we post online — it’s not fully private.”
This connects naturally to broader digital hygiene conversations covered in our article on digital hygiene habits for kids.
What to Watch For Over 3 Months
Month 1: Audit your voice assistant history. Count the false triggers. Delete existing recordings. Set up automatic deletion. Opt out of human review.
Month 2: Evaluate whether smart speakers are positioned in appropriate locations. If a device is in a child’s bedroom, consider relocating it. Test the physical mute process and make it a household habit for sensitive conversations.
Month 3: Revisit what features your household actually uses. Many families discover that a significant fraction of smart speaker functionality could be replaced by phone-based queries without an always-on microphone in the home. The decision to keep or remove is informed by this audit.
Frequently Asked Questions
Are smart speakers legally required to tell children they’re recording?
No current law requires the device to announce that it’s recording. COPPA requires that the company not collect children’s voice data without parental consent, but this operates at the terms-of-service level, not through in-the-moment disclosure. There have been legislative proposals to require audio disclosure, but none have passed as of 2026.
Can someone hack my smart speaker to listen without the wake word?
Yes, this has been demonstrated in research settings. Security researchers have shown that smart speakers can be compromised through various vulnerabilities to enable unauthorized recording. This represents a much smaller risk than the baseline data collection practices of the devices themselves, but it is a real vulnerability that argues for keeping device firmware updated and purchasing from companies with strong security track records.
How is a smart speaker different from a phone with voice activation?
The practical privacy difference is smaller than many people assume. Both your iPhone (Siri) and Android phone (Google Assistant) have always-on wake word detection with similar false-trigger and recording dynamics. The difference is that phones travel with you and have more granular privacy controls in the operating system, while smart speakers are fixed points in your home environment.
Should I remove smart speakers from my home entirely?
That’s a personal decision that depends on how much you value the functionality versus the privacy considerations. The approach taken in this article is to help parents make informed choices with accurate knowledge of what the devices do — not to prescribe removing them. With good settings management, the privacy risk from smart speakers can be meaningfully reduced, though not eliminated.
What about smart TVs that also listen?
Smart TVs with voice activation have the same fundamental dynamic as smart speakers: a microphone that listens for activation. Samsung, LG, and Vizio have all faced scrutiny for their audio data practices. The same principles apply: audit the settings, opt out of voice data collection where possible, and use physical controls (disabling voice activation in TV settings) if you prefer not to have the microphone active.
About the author: Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.
Sources
- Federal Trade Commission & Department of Justice. (2023). FTC Charges Amazon with Violating Children’s Privacy Law by Keeping Kids’ Alexa Voice Recordings Forever and Undermining Parents’ Deletion Requests. https://www.ftc.gov/news-events/news/press-releases/2023/05/ftc-doj-charge-amazon-violating-childrens-privacy-law
- Federal Trade Commission. (2024). COPPA: Children’s Online Privacy Protection Act. https://www.ftc.gov/legal-library/browse/rules/childrens-online-privacy-protection-rule-coppa
- Lau, J., et al. (2018). Alexa, are you listening? Privacy perceptions, concerns and privacy-seeking behaviors with smart speakers. Proceedings of the ACM on Human-Computer Interaction, 2(CSCW). https://doi.org/10.1145/3274371
- Edu, J. S., et al. (2021). Smart home personal assistants: A security and privacy review. ACM Computing Surveys, 53(6). https://doi.org/10.1145/3412383
- Chung, H., et al. (2017). Smart speakers and children: Unintended data collection and use. Privacy Law Scholars Conference. https://privacyscholars.org/
- Amazon Privacy Notice. (2024). Alexa and Alexa Device FAQs. https://www.amazon.com/gp/help/customer/display.html?nodeId=GVP69FUJ48X9DK8V
- Electronic Frontier Foundation. (2024). Smart Speaker Privacy Guide. https://www.eff.org/deeplinks/2019/08/what-should-you-consider-when-using-voice-assistant-home
