AI Needs To Ensure Social Justice In Order To Thrive

Last week I looked at a new report from the innovation charity Nesta, which explored the use of AI technologies in healthcare.  The report highlights the importance of ensuring that any deployment of AI works towards making healthcare accessible and patient focused rather than as a barrier to access.

It’s one of a number of reports that urges those involved in the development of AI to ensure that it progresses in a fair and just manner.  The latest of these comes from a team at the University of Manchester.

The paper argues that any benefits of AI have to be fairly distributed throughout society rather than concentrated in the hands of a few tech giants.

“Ensuring social justice in AI development is essential. AI technologies rely on big data and the use of algorithms, which influence decision-making in public life and on matters such as social welfare, public safety and urban planning,” the authors say.  “In these ‘data-driven’ decision-making processes some social groups may be excluded, either because they lack access to devices necessary to participate or because the selected datasets do not consider the needs, preferences and interests of marginalised and disadvantaged people.”

Making AI fair

The report, which is designed to help employers, regulators and policymakers understand the full effects of AI in a wide range of areas, including healthcare and international policy, was produced by a range of experts from across the University of Manchester.

“Although the challenges that companies and policymakers are facing with respect to AI and robotic systems are similar in many ways, these are two entirely separate technologies – something which is often misunderstood, not just by the general public, but policymakers and employers too. This is something that has to be addressed,” the authors say.

With the tremendous possibilities for AI in healthcare, it’s perhaps not surprising that it’s an area the report focuses extensively on.  They highlight the international work on limiting the spread of antimicrobial resistance (AMR), with the AMR Review providing 27 clear recommendations on how AI can support the work across 10 broad areas.

The report covers a wide range of fields however, and the authors hope that it will add to the growing volume of works that aim to guide and support policymakers, industry stakeholders and regulators better understand the various issues at stake.

Listen to lead author Dr Barbara Ribeiro discuss the report below.

Facebooktwittergoogle_plusredditpinterestlinkedinmail