I buy cheap smart displays to test what manufacturers cut to hit a low price. Over time I’ve learned that “budget” often means compromises you can’t see at first glance—hidden microphones, weak privacy controls, and network behaviour that quietly farms data. If you’re thinking of buying a sub‑£100 smart display (or already have one), here’s the hands‑on checklist I use to spot privacy leaks and detect potential hidden mics. These are practical tests you can run at home, with inexpensive tools and a little patience.
Why test budget smart displays
Cheap smart displays—think low‑cost models from brands like Wyze, third‑party Android displays, or no‑name imports—may offer decent screens and voice assistants, but manufacturers sometimes trim safe design choices or ship software with aggressive cloud behaviour. I’ve seen devices that keep microphones active longer than necessary, send chunks of audio to remote servers by default, or expose sensitive APIs on the local network.
What I look for first: hardware privacy features
Before any network testing, the physical design tells you a lot. I inspect each device for:
Physical mic mute: A true hardware mute cuts power to the microphone. Look for a dedicated switch that’s clearly labelled and tactile.LED indicators: Is there a clear, bright LED that lights when the mic is active? Cheap devices sometimes use tiny or ambiguous lights that are easy to miss.Microphone placement: Spot small pinholes or grille patterns. Multiple holes in odd places can mean multiple mic arrays; note their locations for later verification.Camera shutter: If the device has a camera, is there a manual shutter? A physical cover is meaningful; software-only camera off is weak protection.These checks don’t prove anything, but they guide where to focus software and network tests.
Simple in‑person checks
These are quick actions you can do without tools.
Mute and shout: Flip the mute switch (or press the software mute). Then give the device wake words or commands. If it still responds, the mute isn’t real hardware mute.LED audit: Wake the assistant and watch the LED behaviour. Note whether the LED lights only while responding or remains on between commands—continuous LED suggests microphone is active.Camera/cover test: Close the camera shutter (if there is one) and check whether any video-based features still work. Some systems do degraded behavior when the camera is disabled; others ignore it.Network behaviour: the most revealing tests
I’ve found that network traffic often reveals privacy leaks faster than any visual inspection. You don’t need to be a network engineer—start with these approachable tests.
Isolate the device: Put the smart display on a dedicated guest Wi‑Fi or VLAN. This lets you monitor just that device’s traffic and prevents it from talking to other devices on your LAN.Capture traffic: Use a network monitor such as Wireshark or a simpler router‑based tool (OpenWrt, pfSense, or a Raspberry Pi running tcpdump). Look for continuous outbound connections—these can indicate always‑on audio streams or telemetry.Watch DNS: Cheap devices often hard‑code domain names or use CDNs. Track DNS queries to see where the device resolves hostnames. Repeated DNS requests to an analytics or voice processing domain right after boot or during idle time are suspicious.Check for audio uploads: During idle periods, monitor for large outbound packets or continuous TLS streams over ports 443/5222/etc. If the device uploads audio, you’ll often see periodic chunks of encrypted data leaving the device.Practical tip: run a DNS sinkhole (Pi‑Hole) and watch which queries the device makes. It’s a low‑effort way to see service endpoints without decrypting traffic.
Permission and account settings
Software config is where you can limit damage. I go through these settings right after setup:
Voice recording retention: Can you turn off voice history or automatic review? If the vendor has a dash to opt‑out of human review, use it.Local processing options: Some devices (Google Nest Hub 2, recent Echo models) offer local wake‑word detection and local control for specific features. Budget devices rarely do—if they don’t advertise local processing, assume audio goes to the cloud.App permissions: Inspect the companion app’s permissions. Does it request mic/call logs or contacts unnecessarily? Revoke anything unrelated to core functions.Advanced tests: what I do in the lab
These require more equipment but give conclusive results.
RTFM + teardown (with caution): Read the FCC filings or product manual online for microphone specs. If you’re comfortable, open the device to visually confirm mic hardware. Warning: this voids warranty and can be risky—only do it if you accept those consequences.RF/spectrum scan: Use a handheld RF detector or SDR (Software Defined Radio) to search for unauthorized transmissions. Some rogue devices can transmit low‑power signals when idle. An SDR plus Spectrum‑Viz plugins can show unusual bursts.Audio pump test: Play a constant in‑room noise and check if the device’s network traffic changes when you say the wake word. Correlating audio events with traffic spikes is powerful evidence of audio upload.Man‑in‑the‑middle (for advanced users): Use a TLS proxy (mitmproxy) with custom CA certs if you control the device’s trust store (this is rare on hardened devices). This lets you inspect plaintext API calls to see if audio blobs are being sent.What to do if you find leaks
If a device keeps microphones active or uploads audio continuously, you’ve got options:
Return or replace: The simplest: choose a better model. Devices from reputable vendors (Amazon Echo Show, Google Nest Hub, Lenovo) tend to be more transparent and have stronger privacy controls than cheap imports.Network containment: Keep the device on a guest network with strict firewall rules. Block direct internet access except to vendor domains you trust, or route it through a VPN or proxy you control.Physical mitigation: Use an external cover or position the display so it points away from sensitive areas. If the device has no hardware mute, a cheap hardware hack is a physical microphone blocker (foam plug) placed gently over the mic hole—again, this can be fiddly and may void warranties.Policy opt‑outs: Sign in to the vendor account portal and disable voice review/retention wherever possible. Delete stored recordings and reduce data sharing settings.Quick checklist you can print
| Check | Action |
| Physical mute | Toggle and verify no responses |
| LED indicator | Observe during/after commands for continuity |
| Network traffic | Place on guest Wi‑Fi; monitor DNS and TLS flows |
| Voice retention | Disable history and opt out of human review |
| Local processing | Prefer devices advertising on‑device wake/intent parsing |
Privacy in budget smart displays is a pragmatic risk: many inexpensive models work fine for casual use, but the cheapest often lack meaningful hardware mutes, clear indicators, or robust local processing. If you rely on a smart display in private spaces—bedrooms, home offices, rooms where conversations happen—run at least the simple tests above. They take minutes and can save you from months of unintentional audio leakage.
