RoboTIPS: Developing Responsible Robotics for the Digital Economy

https://www.robotips.co.uk/home

Project Overview

RoboTIPS is an EPSRC Established Career Fellowship led by Marina Jirotka (with Alan Winfield as Fellowship Co-Investigator) that ran from March 2019 to February 2024. The project focuses on transforming social robots into more responsible technologies by addressing a critical challenge: how can we ensure transparency and accountability when accidents inevitably occur?

The Core Problem

Social robots are designed to operate in human environments and interact directly with people. While accidents are hopefully rare, they are inevitable. When incidents do happen, it's vital to understand what caused them so that similar accidents can be prevented in the future. Traditional approaches to investigating robot incidents lack transparency and accountability mechanisms. RoboTIPS addressed this gap through an innovative solution.

The Ethical Black Box

The cornerstone of this project is the Ethical Black Box (EBB), an innovative design feature that fundamentally transforms how we understand and respond to robot incidents. The Ethical Black Box collects data about a robot's actions in real time and in context. When problematic incidents occur, the black box provides a mechanism for the robot to produce a comprehensive account of its own actions. This data enables transparent examination of how an incident occurred, how it may be addressed, and how similar incidents might be prevented in the future.

The EBB serves as a transparency mechanism that enhances user trust in robots by making the robot's decision-making open and accountable. Rather than treating the black box as providing definitive answers, the project positions it as a key part of a reconciliation process, similar to how air accident investigations provide reassurance that helps us continue flying despite inherent risks.

Accident Investigation Framework

A key component of validating the EBB involved designing and conducting a series of experiments simulating robot misbehaviour. Mock accident investigations were conducted with witnesses to reconstruct what happened, producing findings and recommendations in the manner of a real investigatory board or formal inquiry. This approach allowed the team to test whether the Ethical Black Box could effectively support the investigation process and demonstrate its value as an accountability tool.

My Role

I contributed directly to the development and testing of the Ethical Black Box, with a particular focus on the accident investigation component. I helped design and conduct the mock accident investigations, which were critical to validating whether the EBB could effectively support incident analysis. Working with witnesses and reconstructing incident scenarios, I contributed to assessing how the black box data could be used to understand what went wrong, inform recommendations, and ultimately prevent similar incidents in the future. This hands-on involvement with the investigation framework provided insights into the practical challenges of implementing transparency mechanisms in real-world robot accountability scenarios.

The RoboTIPS project shortlisted for the 2025 AI & Robotics Research Awards