Roskachestvo: Identifying Recording and Data Transmission is Extremely Difficult
Smart speakers indeed have access to a significant amount of user information. For instance, the device *could* potentially send conversation data, even to malicious actors, although this is considered highly improbable. This was reported by the Roskachestvo Digital Expertise Center, which also admitted that identifying specific user activity, such as recording a conversation and transmitting it, is extremely challenging.
The most widespread and concerning question is whether these gadgets “eavesdrop” and transmit all user conversations. While a speaker remains in constant listening mode, waiting for an activation phrase (like `Alice` or `Marusya`), stating that it records and transmits everything would be incorrect, the organization clarified.
Interaction between the device and the server doesn`t occur as frequently as one might assume, they added. Typically, network identifiers are sent to ensure the network confirms the device`s operational status and connectivity.
“Identifying specific activity, such as recording and sending a conversation, is extremely difficult. This is due to the use of complex encryption algorithms designed to protect user confidentiality. Consequently, the question of whether data from overheard conversations is transmitted remains open,” Roskachestvo representatives noted.
Another common fear, according to their data, is that hackers might compromise speakers to listen in on private conversations. While theoretically possible to hack a smart speaker and use it as a “bug,” experts question the practicality of such actions, as everyday conversations are unlikely to hold significant value for cybercriminals.
The assertion that smart speakers collect personal data for targeted advertising is partly true, but with certain reservations. Everything a user says near a smart speaker *can* indeed be used to create a detailed profile of interests, habits, and even routines. By accepting the smart speaker`s privacy policy regarding data collection, the user consents to such data processing, but in this scenario, the advertiser typically receives the information in an anonymized form.
Users also often worry that malicious actors might record their voice through speakers for identity theft and fraud. While theoretically possible to use a voice recording obtained via a smart speaker for identity theft, this scenario is practically improbable. Fraudsters find it much simpler to use readily available online recordings of public figures to create a higher quality deepfake.
Sergei Kuzmenko, head of the Roskachestvo Digital Expertise Center, stated that while complete digital security is unattainable, understanding risks and continuously increasing awareness significantly reduce the likelihood of becoming a victim of cybercriminals.
According to statistical data from analysts at M.Video — Eldorado, smart speaker sales in Russia grew by 33.3 percent in the first quarter of 2024 compared to the same period in 2023, exceeding 1 million devices sold.
Rupert Blackwood 
Barnaby Whitfield