The commercial web is built on business models that aggregate personal data for the benefit of advertisers. What would the web look like if consumers had control over the algorithms that process their data?

Develop a system that ingests a user's loyalty-card or banking data. Use it to feed a recommender system that interfaces with a comparison shopping API to present the user with personalised offers for products they are likely to need, according to preferences they set, without providing any personal data to any third party.

One of the classic trade-offs on the web is that users give up control of their data in exchange for the convenience and benefit they receive from personalised services. Imagine instead that a user could keep their data securely within a personal data store and use private computing containers to run AI systems that process their data, all fully within their control. The ideal system would enable users to receive the benefits of personalisation that works in their interests, without providing their personal data to third parties. The system should be sensitive to a user's revealed preferences (their history) as well as to their express intent. For example, a user may opt to prefer goods or services that are more environmentally friendly, rather than being presented with offers that maximise revenue for a centralised commercial platform. In the long term, personalised, de-centralised computing provides an opportunity to build a more open, diverse web that works in the interests of users.

Examples

  • As a privacy-conscious consumer, I want to be presented with relevant offers that I can trust, so that I can get the convenience of Amazon's recommendations without handing over my data or submitting to their relevance and ranking criteria.
  • As a consumer who buys clothing online, I want to be presented with relevant offers for sustainable, ethically produced goods, so that I can conveniently make more ethical choices.

Potential resources: