Trustworthiness is essential to the uptake of digital technologies. The European Commission’s (EC) recent whitepapers on Europe’s data and AI strategies are littered with policies intended to enable consumer trust in the organisations and infrastructure that collect and process data about them.
This approach views trust as something that can be established through compliance with security and privacy regulations. That form of trust is essential: consumers need to be confident that the laws and regulations put in place to protect them are adhered to, but reciprocity is also essential for a sustained, trusting relationship between consumers and businesses.
Too many digital business models lack perception of true reciprocity. Hence, the increasingly popular saying – If you aren’t paying for the product, you are the product. It is easy to see how collecting granular user data helps enterprises drive revenues by microtargetting advertising and pricing, or to optimise internal operational decisions.
It is often harder to appreciate how data-enabled personalisation might benefit the consumer – and that is largely down to a lack of transparency to users.
In short, the current dominant approach is for digital businesses to expect that the consumer is transparent to them whilst offering relatively little visibility in return. It is very much like a one-way mirror. Whether it is home voice assistants taking audio snippets to the cloud or wearables sending streams of biometric data; the ‘consumers’ data’ is often not readily available for review or transfer to a competing service. This means the benefit of that data is often skewed in favour of the business over the consumer – leaving the perception that it is simply used to microtarget advertising and products down to the individual and extract the best possible RoI.
Now there is perhaps some consumer benefit – finding the product you want faster – but are you finding the product you want, or the product Amazon wants to sell you? How can you know the difference?
The EC’s recent data and AI strategy do not address this. Pooling data centrally for the benefit of any European company risks leaving consumers behind in their understanding of how their data is being used, generating a lack of transparency and fuelling the perception that their data is being used against them.
Instead, to deliver an ecosystem of trust, the EC should be looking for ways to remove the one-way mirror and allow users to take control of their own data and share it much more selectively, and for a more equitable set of benefits. Below is an outline of a few contenders:
- Blockchain provides a mechanism to retain data-ownership which in turn could allow users to gain control, or financial benefit from the way in which their data is being used. Unfortunately, distributed ledger approaches can be energy intensive, presenting a major challenge as data volumes increase and the world’s need to improve energy efficiency, in the face of climate change, becomes more pressing.
- Federated learning is another technology that could form part of the solution. Rather than pooling data centrally in order to apply machine learning techniques, federated learning instead learns at the edge. A machine learning model is deployed on the end user’s device, or at their premises, and it sends only what it has ‘learnt’ back to a central node in the cloud. Each of those learnings is combined centrally and shared back out to the user’s devices or premises for use. With this approach, private data never moves across the network, and there is the potential to add filters that reflect users’ personal preferences around privacy trade-offs. Eventually, users could build up a personal digital twin, and allow access to relevant parts only to trusted businesses.
- APIs offer a potentially revolutionary approach. The Open Banking movement which relies on this approach provides an analogue. The Open Banking model allows users to share the benefit of data co-created with their bank for the benefit of other financial relationships or transactions. Crucially user data doesn’t become siloed for the exclusive benefit of one firm, spurring innovation and competition over oligopoly.
We should envisage an ecosystem of radical transparency where enterprises open up their data and algorithms, to allow users to explore and customise products and services for their own needs. This would be a digital economy characterised by accessibility rather than personalisation, where the benefits of data and AI act for mutual benefit.
Consumers and enterprises naturally distrust being placed in a position of asymmetric power. For an ecosystem of trust to emerge, digital leaders will need to architect reciprocal business models where data is co-created and used for mutual benefit. Rather than thinking in terms of microtargeting, we all need to think in terms of empowering and enabling consumers.