Straight to content
Image: Andreas Dantz (CC)

European Commission will investigate in-depth Amazon’s takeover of iRobot

Posted in category:
Long read
Written by:
Written by: Margarida Silva
Published on:
reading time 5 minutes

On 6 July, the European Commission’s competition department announced(opens in new window)  its in-depth investigation into Amazon’s plans to buy home robotics company, iRobot. The Commission’s decision echoes SOMO’s concerns that the deal would entrench Amazon’s market dominance at the expense of people’s privacy and livelihoods.

SOMO, Foxglove(opens in new window) the Balanced Economy Project(opens in new window)  and the Open Markets Institute(opens in new window)  in February had asked the European Commission to thoroughly scrutinise the deal as we are concerned that by buying the Roomba maker, the tech giant would be getting eyes inside our homes and, yet again, boosting its already immense market power.

The Commission’s decision to deepen its investigation was based on its concerns that the deal would enable Amazon to:

The first concern follows the well-tread territory of analysing potential blocking, or limiting, access by rival companies, often called foreclosure risks. This is an important consideration as Amazon already is(opens in new window)  one of the dominant players in the home connected devices market (both selling its own products and as a sales channel) and it controls one of the main gateways for people to interact with different devices, Alexa.

Yet, arguably more important for this case is the role of home data.

What are iRobot’s data?

The NGO  Privacy International(opens in new window)  has put iRobot’s Roomba to the test to understand exactly what type of data it is collecting and how it could be used by Amazon. The team bought and used a WIFI Connected Roomba® i7 Robot Vacuum, then analysed its Privacy Policy and the information received via Data Subject Requests.

The results are enlightening. iRobot’s vacuums collect a variety of data on users and their home environment, including personal data which it directly receives from users (name, address, billing information, serial number) or from third parties (i.e. other apps or social networks use to log in) and  data collected directly by the Roomba. The latter is particularly important as it can include things such as floor plans, images and video, labeling of furniture, and detection of other connected devices.

Screenshots from the iRobot Home app interface showing, among other things, the level of detail about users’ homes iRobot may have access to, including the existence of other connected devices by identifying their MAC addresses.

Such detailed and intimate home data can be extremely valuable especially if it is used to profile and make predictions about people. As we are dealing with home data, the inferences can be incredibly personal and sensitive including a persons’ financial status, health, religion, and purchasing habits.

Why does it matter?

By taking over iRobot, Amazon is not only getting a home robotics company, but a home data treasure trove that it can pair up with its already immense amount of consumer data. And the European Commission’s decision recognises that. In the press release, the Commission explained(opens in new window)  that iRobot’s user data could “provide Amazon with an important advantage in the market for online marketplace services to third-party sellers (and related advertising services) and/or other data-related markets”.

Amazon can then use its data processing capabilities and different AI tools – including machine learning – to understand and use that data to benefit other parts of its business: including retail, marketplace services for independent sellers and advertising.

iRobot’s data, when combined with Amazon’s own data and processed, could be used to profile consumers and target them at vulnerable moments, identify potential new customers for its own products and push personalised ads to them. Privacy International offers a few examples of how this could work:

Amazon would be in a position to know that an iRobot user has bought a new smart TV and suggest accessories for the TV or even an Amazon Prime Video subscription. Another example would be to identify, for instance, a small dining table added by the user and the free space around it, and then serve the user with ads for a bigger table.”

Data advantages, advertising and AI

Amazon already uses data in a wide number of ways: from automated decisions on whether to start or end the sales of a product, set prices, manage inventory and to target advertising. The European Commission’s announcement highlights two potential concerning uses: marketplace intermediation services for third-party sellers and advertising.

Amazon is, by far the biggest marketplace in Europe. SOMO’s research has shown how Amazon already holds monopoly power over third-party sellers in Germany, the UK, France, Spain and Italy. This has allowed the company to get away with extracting ever higher fees from sellers, pushing them into paying for extra services and using their confidential data to compete against them.

The company’s power over sellers has also enabled it to monetise its product search by replacing organic results and features with sponsored content.

In 2021, Amazon gained EUR5.4 billion in advertising revenue. A whole 17 times as much as it had in 2017.

The European Commission is concerned that, by gaining access to iRobot’s data, Amazon could, for instance, further develop its search results and/or deepen the personalisation of sponsored content. This would make sellers even more dependent on the platform and would allow Amazon to charge even higher fees – especially in advertising – from sellers.

Ultimately, if correct, this could lead to independent sellers being pushed out of the market entirely, and to higher prices for consumers.

Home data to train machine learning

iRobot’s data could also be used by Amazon to train its artificial intelligence (AI) tools.

In the past months, there has been growing public interest in AI techniques, especially generative AI, and their potential impacts. Yet, one often overlooked aspect of the developing field  is how it has so far been controlled(opens in new window)  by Big Tech firms such as Microsoft, Alphabet (Google) and Amazon.

AI techniques, especially machine learning, are already deeply embedded in Amazon’s business. As Amazon Web Services’ vice-president of Data and Machine Learning Swami Sivasubramanian explained(opens in new window) :

Our e-commerce recommendations engine is driven by ML [machine learning – ed.]; the paths that optimize robotic picking routes in our fulfillment centers are driven by ML; and our supply chain, forecasting, and capacity planning are informed by ML. Prime Air (our drones) and the computer vision technology in Amazon Go (our physical retail experience that lets consumers select items off a shelf and leave the store without having to formally check out) use deep learning. Alexa, powered by more than 30 different ML systems, helps customers billions of times each week to manage smart homes, shop, get information and entertainment, and more.”

Beyond using AI in its own operations, Amazon also sells these products to business and government clients via its cloud services.

The fact that such a supposedly disruptive technology is dominated by Big Tech firms like Amazon, stems in part from the huge amounts of data it has access to, either through its own business operations or via mergers and acquisitions. As the AI Institute Now(opens in new window)  has put it:

Firms that have access to the widest and deepest swath of behavioral data insights through surveillance will have an edge in the creation of consumer AI products. This is reflected in the acquisition strategies adopted by tech companies, which have of late focused on expanding this data advantage.”

Amazon’s acquisition of iRobot – itself a company with(opens in new window)  expertise in machine learning – and access to its trove of home consumer data, can further propel Amazon’s dominance in AI. This, in turn, can entrench its power in a variety of sectors.

Welcome decision to start a Phase II investigation

The European Commission’s decision to further investigate this deal and, in specific to dig into its data dimension, is a particularly welcome development. Some commentators – and even competition authorities – have underestimated the importance and the risk raised by this deal in part because they have not fully considered the data dimension.

The UK’s Competition and Markets Authority (CMA), for instance, recently approved(opens in new window)  the deal after an investigation that narrowly focused on the potential impacts to the robotics vacuum cleaner market.

The Commission’s announcement shows a move in the right direction. Now, the Commission has 90 days to gather more evidence, hear from third-parties and discuss its concerns and possible remedies with Amazon and iRobot.

In past cases however, the European Commission has approved deals even after conducting thorough investigations that confirms concerns. In the controversial Google-Fitbit(opens in new window)  case, the Commission allowed the takeover to go ahead once Google committed to keeping data in siloes. However, such remedies that rely on corporate giants to change their behaviour are difficult to monitor and even harder to enforce.

We hope that during its investigation, the Commission will take into account this experience and, if it confirms its concerns, block the deal.

Posted in category:
Long read
Written by:
Written by: Margarida Silva
Published on:

Related news

Don't want to miss anything?

Sign up for our newsletter and always stay up to date on information and analysis on corporate power issues.