Alice Enders and Joseph Evans ask if voice-activated speakers threaten the privacy of our homes
Smart speakers such as the Amazon Echo and Google Home accelerated their prodigious rate of adoption in the UK in the last quarter of 2018. Forecasts suggest that they will shortly be in a fifth of British homes.
But this trend of rising adoption could soon hit a wall. Surveys suggest that the majority of Britons fear that their privacy may be compromised if they invite voice-activated speakers into their homes.
The context is the omnipresent nature of voice interfaces on smart devices. The virtual assistants that power voice interfaces, such as Apple’s Siri and Amazon’s Alexa, are used by many Britons to operate their smartphones and tablets. Always at our sides, these devices are also frequently listening out for their wake word and, so far, without much of a consumer backlash over privacy. However, while most people find a “hands-free” option a useful way to operate these devices, smart speakers are in a class apart. Unlike the smartphone, they are a pointless purchase if you don’t feel comfortable about using a voice assistant. The logic of a smart speaker is that it is left constantly on, the voice assistant primed to receive its wake word – “Alexa” in the case of Amazon’s Echo.
This is not the same as an always-on device that records the chatter amid the clatter of dishes in the kitchen, say, to be stored on a server to haunt the user with dark secrets gained by eavesdropping. Even so, high-profile bugs and false positives, where conversations were recorded by mistake, have not helped allay the public’s fears.
Currently, Amazon is far ahead of Google in the number of homes served by its speaker. At three times the price of Echo, Apple, meanwhile, has never sought to win over the mass of homes with its HomePod speaker. However, for privacy-conscious customers, its speaker is more suitable. It does much more of its processing within the device or on a paired iPhone, limiting the upload of conversations to the Cloud.
On the face of it, there seems to be no commercial advantage to a company such as Amazon in recording all conversations – they might only terminally confuse poor Alexa – instead of sticking to the plot of listening for wake words to record instructions, transmit them to the server and promptly fulfil them.
By pushing customer adoption of its platform, Amazon is looking to future-proof its business through data collection and by controlling a so-called discovery layer between customers and services.
Google has a similar objective for its low-priced smart speaker, Home – but Amazon Prime’s 10 million UK homes give it a huge marketing advantage.
Google is interested in smart home devices (it bought the smart thermostat maker Nest), but a future income stream from Home is less obvious than for Amazon, since Google’s core business is search advertising, rather than families buying goods and services.
Currently, data collected via Echo is probably Amazon’s single most important benefit, because the data can be used for so many different ends. The company can learn more about customers’ demographics, interests, habits and tastes from the following: the requests people make for information, such as “What is today’s weather forecast?”; the times of day that they are at home; the number of people in a household; people’s media preferences, including their favourite music and news choices; and personal contacts, if the user connects their phone to the speaker.
This data could eventually support a richer experience than just responding to instructions. Machine learning-based intelligence will extend beyond translating natural language and images to machine-readable information, and start to deliver recommendations, suggestions, reminders and notifications.
One again, data is the key driver for Google, as it can feed voice requests into its enormous data engine and so improve the effectiveness and targeting of its adverts.
Although the motivation is similar for both companies, Google is, if anything, more exposed to customer suspicion simply because its status as a business driven by data is better understood by the public. It therefore needs to be unimpeachable on the privacy front. Scandals such as the early versions of the Home Mini recording without being woken up are exactly the sort of thing it needs to avoid.
Voice-activated speakers will be able to do more than respond to commands to magic up a radio station, answer a homework question, play this show or text that person. They will be capable of responding to more sophisticated instructions, such as: find me something to watch tonight, book me a table for lunch, or tell me what’s going on in Parliament. In other words, Alexa and her cousins will be capable of making personalised decisions on our behalf.
"The strategic goal of the smart speaker’s supplier is to be the gatekeeper to the home"
The key strategic goal of the smart speaker’s supplier is to be the gatekeeper to the home. The lower it prices its devices (such as speakers or tablets), the more people will buy them, giving Amazon a foothold in the home that is a gateway to its storefront, including for media. In this respect, Amazon treats its devices as wedges that open up the home to those of its services where it makes the real money.
To the extent that Echo owners ask Alexa to select or recommend video or audio content, Amazon gets to influence the process. Amazon could choose to prioritise its own original content or sell access to other content providers. We suspect that few Echo owners are currently asking Alexa for TV recommendations, but this is part of the long-term vision for voice user interfaces.
Third-party service providers and app developers promise a rich and varied suite of services to users. For the moment, though, Amazon is not charging third-party services for access to Echo users.
This makes sense for now: the aim is to get the speaker into as many homes as possible and for it to be used as widely as possible. Eventually, affiliate fees from third parties should be a fairly flexible way for Amazon to extract value from its position between businesses and customers.
But, in the short term, data collection seems to be the primary commercial aim. Amazon’s strength here is enhanced by the billing relationship it already has with every Echo owner via their Amazon accounts.
Amazon and Google have built these devices to gain entry to the home, and they compete against each other by offering low prices – to the benefit of purchasers. The revenue potential of a powerful platform position inside the home is only likely to be realised in the future. And it will require the user to be ever more intensively engaged with the device, to the point where it becomes used for many purposes. At that point, a user will be communicating a great deal of information about who they are and what they like to Alexa and Echo.
Consumers are increasingly aware that they are often the product, not the customer, in the online economy – they pick up on the steady stream of data breaches, hacks and scandals, and are understandably nervous about installing a connected microphone in their homes. The abuse of Facebook personal data by Cambridge Analytica is only the most prominent of these scandals.
"Amazon has a pressing need to build trust"
Although Britons are not as touchy as Germans, say, when it comes to privacy, they are certainly more alert than US internet users, for whom Amazon first developed the Echo.
We don’t expect the average Brit to acquire all things new and shiny just because they are labelled “smart”. Amazon has a pressing need to build trust, because the present absence of trust is a barrier to Echo’s success. The company does inform users on data policy, as part of its obligations under the General Data Protection Regulation that came into force in May 2018 – but compliance is in its early days.
Much remains to be clarified for potential owners of these devices. We expect the UK regulator, the Information Commissioner’s Office, to contribute to reversing the climate of mistrust by pursuing compliance with core principles that apply to the gathering and processing of personal data.
Alice Enders and Joseph Evans are analysts at Enders Analysis.