Pet robots and psychological harms

Dress rehearsal, 03 April 2023 – Responsible Technology Institute, Oxford

Official event, 21 June 2023 –  Responsible Technology Institute, Oxford

A challenging scenario for the Ethical Black Box and robot investigators.

June 21st, 2023 - Responsible Technology Institute (Oxford). We have successfully run the second of our mock robot accidents and investigations. This time, we analysed the role of the Ethical Black Box in the reconstruction of an incident involving a young girl and her toy robot. 

The scenario was about the emotional attachment between a young girl and her robotic dog, which becomes addictive and a source of stress due to several causes, including a software virus that changes the robot’s behaviour.

For ethical as well as practical reasons we did not simulate the accident and its consequences on the girl (i.e., psychological harm) but only the investigation process. By collecting witnesses’ testimonies and inspecting the data recorded by the Ethical Black Box, investigators were able to determine the cause of the robot misbehaviour. This investigation highlights both the usefulness of the Ethical Black Box in examining human/robot accidents, and the risks related to robot anthropomorphism, in particular for vulnerable people such as children.

We are now preparing the next simulated accident and investigation, which will focus on bionic prostheses.

Huge thanks to all participants in the dress rehearsal and official simulation event: (in alphabetical order) Paul Bremner, Sam Brown, Ken Fairbank, Benjamin Hardin, Chris Harper, Zsofia Lazar, Carl Macrae, Jaime Stewart, Carolyn Ten Holter, Tyler Reinmund, and Helena Webb. 

The RoboTIPS project as a case study of participatory research

For more information please visit Participatory Research Oxford:

First mock accident investigation. Done!

A special thanks to the volunteers for participating in our RoboTIPS mock accident and investigation.

July 15th, 2022 - Bristol Robotics Lab (Bristol). RoboTIPS team is proud to announce that we have successfully staged the full mock investigation involving an accident with a care robot incorporating an Ethical Black Box. The accident scene was brilliantly enacted by 5 non-professional volunteers alongside a Pepper robot, programmed with a predefined behaviour. Two investigators – experts in the fields of accident investigation and robot safety, respectively – were given the task of uncovering the sequence of events by inspecting the scene of the accident, questioning the witnesses and examining the robot’s data logs. In the end, the investigators correctly deduced the role that the robot played in the accident, providing helpful feedback on how the EBB can be used in practice. We are immensely grateful to our volunteers who contributed to a very successful event!

As to the next steps, we are currently drafting a full report on this mock accident investigation, which will be published soon in an open access journal. Moreover, we are now preparing a new robot accident/incident scenario for the next investigation!  

Andrew and Kathleen Booth Memorial Lecture

15 June 2022 - London. Professor Marina Jirotka delivered the 2022 Memorial Lecture, which was an opportunity to celebrate the life and work of Kathleen Booth. In her lecture, Professor Jirotka talked about the RoboTIPS project and discussed the importance of maintaining public trust in new technologies. More information available here.

Above: The scene of the accident. Below: one investigator collecting witnesses' testimonies.  

First mock accident investigation: a successful pilot

April 7th, 2022 - Bristol Robotics Lab. RoboTIPS team ran a valuable pilot test of the mock accident investigation involving a social care robot (Pepper). The goal of the pilot was to rehearse the accident scenario and the investigation procedure in preparation for the full event, which will take place during the summer. The mock accident investigation is one of the research objectives of the RoboTIPS project, in which we aim at answering the following questions: (1) What does an investigation process for robot-related accidents look like? (2) What role does/could black box data from the robot play in this process? And (3) To what extent do human-robot interactions need to be logged, and how, in order to satisfactorily inform the accident investigation process? The pilot mock accident investigation went very well - the investigators correctly deduced the role that the robot played in the accident. For a more detailed account, you can visit Alan Winfield's blog: