Building Public Value via Intelligible AI

The PubVIA project aims to build public value via intelligible AI (Artificial Intelligence) and to enable the BBC to consider the potential threats and risks AI and IoT (Internet of Things) pose for journalism. By using the insights and knowledge generated from two years of embedded research in the BBC’s editorial, product and policy teams on the PETRAS ICE-AI project, PubVIA’s objective is to impact how these systems are made understandable and usable for journalists.
News organisations are increasingly using AI and connected IoT devices to gather, produce, and distribute their news content. There are voice assistants reading the headlines, automated tools writing stories from data, and algorithmic recommenders selecting personalised news agendas. That’s not to mention behind-the-scenes development of synthetic media, ‘immersive’ reporting, and sensors for the living room. However, this data-driven innovation raises urgent and difficult challenges including cybersecurity issues, risk of digital harms, as well as questions of ethics, trust, and understanding. Public service media such as the BBC are taking these issues seriously and looking for evidence-based ways to foresee the consequences of these new technologies and forestall the negative impacts they could have.
Pulling together an experienced team from the University of Edinburgh of experts in speculative design, ethics and journalistic practice, the PubVIA project aims to achieve impact at the BBC through the following activities: 1) Co-designing and delivering six workshops by using a range of ‘future scenarios’ of AI and IoT to prompt knowledge exchange between technologists and journalists, 2) Working with policy and strategy teams to inform guidelines for intelligible, ethical and responsible AI and IoT, 3) Working with BBC design teams to tailor resources for engagement and education of practitioners, including a store of context-sensitive scenarios that can be used by others.