Research

Building Public Value via Intelligible AI (PubVIA)

Ewa Luger, Chris Elsden, Bronwyn Jones

In collaboration with:

EPSRC

1st October 2022 – 31st March 2023

Aim 

The primary goal of the PubVIA project was to optimise public value through comprehensible AI (Artificial Intelligence), helping the BBC navigate potential AI and IoT (Internet of Things) hazards for journalism. Drawing on the insights generated from two years of immersive research within the BBC’s editorial, product, and policy teams through the PETRAS ICE-AI project, PubVIA aspires to influence the clarity and accessibility of these systems for journalists. 

Rationale 

News organisations are progressively utilising AI and IoT-connected devices to collect, create, and distribute news. Voice assistants announcing headlines, automated tools creating narratives from data, and algorithm-driven recommenders curating personalised news are becoming the norm. Add to that the backstage development of synthetic media, ‘immersive’ reporting, and living-room sensors. However, these data-centric innovations present pressing issues, including cybersecurity issues, risk of digital harms, as well as questions of ethics, trust, and understanding. Public service media organisations like the BBC are earnestly addressing these matters and seeking evidence-based strategies to anticipate the implications of these emergent technologies and preempt their potential negative repercussions. 

Approach 

The PubVIA project, led by a seasoned team from the University of Edinburgh, specialising in speculative design, ethics, and journalistic practice, plans to make impactful contributions at the BBC through several activities: 

  1. Co-design and host six workshops using an array of ‘future scenarios’ of AI and IoT to facilitate knowledge sharing between technologists and journalists. 
  2. Collaborate with policy and strategy teams to develop guidelines for intelligible, ethical, and responsible AI and IoT usage. 
  3. Work with BBC’s design teams to customise resources for practitioner engagement and education, including a repository of context-sensitive scenarios for wider use. 

 

For more information:

Petras – – Building Public Value via Intelligible AI (PubVIA) (petras-iot.org)