The University of Southampton

HCI and AI combine for new, meaningful model to Data Sharing

Published: 15 December 2017
Illustration
ACM Interactions Cover Story

With the rise of the Internet of Things (IoT) it’s possible to conceive that every fridge could know your name and what you eat to be shared with a plethora of service providers without you actually giving your consent. As the Internet of Things grows, more and more devices in our homes and in our environments will be gathering data about us: touch a door to a building, we’re monitored; move through a subway or on a train or enter a shop, we’re being tracked. It’s not just web pages following where we go anymore to serve us ads. With the increase in scale and speed of requests for our data, we need new ways to manage these requests – and to ensure that those requests treat our data appropriately. The current paradigm of ‘see a service, lie that we have read the terms and conditions each time we hit “accept”’ won’t scale.

But what if it were possible to make consent both more meaningful and less disruptive than the current meaningless terms and conditions button-pushing?

ThatÂ’s where the research of University of Southampton Computer Scientists led by m.c. schraefel, Professor of Computer Science and Human Performance, becomes extremely useful, developing concepts to envision and shape the IoT as a consentful, human-centred space.

Professor schraefelÂ’s latest research features on the cover of the current issue of Interactions, the flagship magazine for the Association of Computing Machinery (ACM)Â’s Special Interest Group on Human-Computer Interaction (HCI). Working with Southampton colleagues Associate Professor, Dr Enrico Gerding, Dr Alper Alan and PhD researcher Richard Gomer, and Carsten Maple, Professor of Cyber Systems Engineering at Warwick, Professor schraefel and her team propose a radical new approach to personal data sharing that combines AI and HCI.

“Our aim is to help empower citizens and services to have a better more informed relationship about use of personal data without having to deal with these decisions constantly, personally,” says Professor schraefel. “We want to design interactions between AI and a person both to capture a person’s data sharing preferences, so that agents can negotiate with a service about the terms and conditions of data access, as well as help track that use and revise it.” Schraefel compares the approach to reviewing a bank statement with a trusted advisor: “The current model has us asking for access when our focus is on something else; terms are also all one way for the service’s benefit. HCI with AI can change this dynamic to deliver a new model of benefit to all. How we do that well is the subject of our research.”

Click here to read the full article ‘The Internet of Things: Interaction Challenges to Meaningful Consent at Scale’ in Interactions.

Articles that may also interest you

Share this article FacebookTwitterWeibo