For the past two years our UnBias project has been exploring issues around ‘unfairness’ in algorithmic systems. Our research work has identified the existence of concerns across different societal and professional groups over the contemporary prevalence of algorithm driven online platforms. At the same time, we have also identified amongst these groups a desire for change to improve the user experience of platforms. Our analysis highlights several opportunities for positive change – in particular in relation to education, societal engagement and policy.
As the UnBias project draws to a close, we can a report a range of substantial findings. A summary can be downloaded here or watched as a short narrated slideshow below.
You can also find more details about our key findings and outputs here.
From December 2018 our project team will be working on a new project to follow on from UnBias. ReEnTrust will identify mechanisms to foster trust between users, algorithms and platforms.
For two years the UnBias project has been examining the user experience of algorithm driven internet platforms and seeking answers to important questions such as:
Are algorithms ever neutral?
How might algorithmic systems produce unexpected outcomes that systematically disadvantage individuals, groups or communities?
How can we make sure that algorithmic processes operate in our best interests?
On October 1st 2018 we held our UnBias project showcase event. Seventy attendees from the fields of research, policy, law, industry and education came together at the Digital Catapult in London to hear and discuss our key project findings. We also highlighted our practical outputs such as policy guidelines and our exciting fairness toolkit.
The event also included a lively panel debate and engaging presentations from external speakers about the social consequences of algorithmic biases and how they might be addressed. We ended the day by announcing plans for our follow-on project, ReEnTrust, which will identify mechanisms to rebuild and enhance trust in algorithmic systems.
Watch an overview of the day here!
Emancipating Users Against Algorithmic Biases for a Trusted Digital Economy