We talk to Andrew Burgess (CEO of Greenhouse Intelligence, AI strategist and management consultant) about how he was able to apply a new data ethics lens to his work after taking the Open Data Institute (ODI)’s Data Ethics Professionals course.
Andrew Burgess, CEO of Greenhouse Intelligence, is an AI strategist and management consultant, as well as author of the Executive Guide to Artificial Intelligence, visiting senior fellow Loughborough University and Expert-in-Residence for AI at Imperial College Enterprise Lab. He is also an AI advisor to the Natural History Museum and STEM Ambassador. He creates AI frameworks for businesses and helps them to implement their strategies, and has worked with various charities including Macmillan, Prostate Cancer UK and Parkinsons UK.
In his consultancy work, Andrew helps businesses, who are often in the early stages of their AI journey, to identify their objectives and understand how AI can be embedded into their organisations in order to help them reach their goals. As part of this journey he wanted to also be able to show how to incorporate data ethics when applying AI technologies to minimise any potential harms arising as a result of collecting, using or sharing data, and was looking for a formal qualification to help him.
Working with the ODI
Andrew attended the ODI’s Data Ethics Professionals course. Having a professional qualification meant that Andrew was able to confidently deliver training in the Data Ethics Canvas to clients, providing a practical framework for companies to work from. Since then, he has also become an assessor for the programme, and has delivered demonstrations for facilitators. He has also created a webinar for Dataiku on real-world AI ethics which had over 250 registrations, and is working with the ODI on further projects.
Working with the Data Ethics Canvas and learning how to apply it to various scenarios has enabled Andrew to retrospectively examine other projects that he has worked on.
One example is Parkinson’s UK – the largest charity funder of Parkinson's research in Europe. During 2020, the Covid pandemic meant that ‘wet’ lab research into Parkinson’s could not be carried out, so Andrew helped the organisation plan remote, AI-based research, using access to several longitudinal cohort studies funded by Parkinson’s UK and the Michael J Fox Foundation in the USA.
The research identified three sub-types of Parkinson’s, using common biomarkers which can be measured directly such as gait, tremor, cognition and sleep behaviours. By accurately identifying which sub-types a patient falls into, medical practitioners can gain a more accurate prognosis of a patient’s disease and enable more targeted treatments and better outcomes.
Parkinson’s UK are currently seeking funding to build a clinical tool for use by GPs and hospitals. This front-line tool could provide a fast and accurate method to predict the sub-type of Parkinson’s a patient is likely to be suffering from at an earlier stage, enabling more effective care pathways with more specific therapies and medication.
However, reviewing the data through the lens of the Data Ethics Canvas showed that cohorts in the anonymised studies were male-dominated (more so than can just be taken into account by the natural bias towards Parkinson’s being diagnosed in men) and that less people from ethnic minorities were included.
With these revelations, recommendations can be made that steps are taken to address these anomalies when looking at the next cohort of participants, and the canvas can be used to take these and other considerations into account for future trials to make sure they are properly representative – for example by seeking out additional data sources.
"There are huge amounts of latent value in an organisation’s data, which AI has the power to unlock, and I aim to help companies become more competitive and innovative by developing effective AI strategies. However data, and data ethics, are vital components in this journey and must be taken into account in order to gain and retain trust, so taking the Data Ethics Professionals course has been very useful in giving me an understanding of how to do that using an accepted practical approach which can be applied in a range of use cases."
The ODI is building a network of Data Ethics Professionals and Facilitators around the world.
This network of individuals have demonstrated themselves as competent in data ethics and the practical steps required to help organisations minimise potential harms that come from data collection, use and sharing.
‘ODI Data Ethics Professional’ and ‘ODI Data Ethics Facilitator’ is a quality mark for those who have been assessed by the ODI through our Data Ethics Professionals and Facilitators programme. Certified individuals listed here can deliver professional services to organisations under the ODI Data Ethics Professionals and (where applicable) Facilitators badge.