A report that outlines proposed rules for governing artificial intelligence (AI) has today been published by the European Commission, after consultation with researchers from across Europe, including De Montfort University Leicester (DMU).
Expert academics from DMU who lead the European Union-funded SHERPA project, which aims to enhance the responsible development and use of AI and examine the impact on ethics and human rights, have been consulting with and providing recommendations to the European Commission officials preparing the report.
(Image: Possessed Photography/Unsplash)
The draft regulation, which includes a ban on surveillance, strict safeguarding rules for high-risk AI applications and hefty fines for violations, is thought to be an important step towards a European regulatory framework that will govern AI in EU member states.
A debate involving the European Parliament and all 27 member states will now commence, before a definitive set of regulations is agreed and put in place.
“Regulating AI is a topic of lively debate within various communities, including technical computer science and AI research communities, but also other academic communities such as law and social science,” said Professor Bernd Stahl, director of the Centre for Computing and Social Responsibility at DMU and project coordinator for SHERPA.
“It is also of high interest to industry which sees significant potential to benefit from AI, as well as civil society groups who are worried about possible consequences and policymakers who will seek to ensure economic and other benefits of AI can be realised, while problems and downsides are addressed.”
As part of the SHERPA project, Professor Stahl and his colleagues have conducted significant research and stakeholder consultation to find out how AI technologies are perceived in practice, which ethical and social issues people are concerned about, and how these can be addressed.
“We welcome the European Commission's proposed regulation on the use of AI, as we agree that a regulatory framework will be required to provide certainty to AI developers, deployers and users,” continued Professor Stahl.
RELATED NEWS
Researchers advise European Parliament on responsible AI development
DMU to lead Europe-wide computer ethics project
Researchers explore how artificial intelligence could impact our lives by 2025
The draft proposes a number of measures to strengthen AI research and innovation which will be required to allow European AI to retain some independence from US and Chinese competitors.
It introduces a risk management approach for high-risk AI systems, including those that are used for remote biometric identification of people in public spaces such as large-scale facial recognition systems.
It also sets out that developers will be required to conduct detailed self-assessment on high-risk systems such as those that determine access to education and AI used for human resources purposes including filtering of applications or evaluation of candidates.
Other examples include financial systems that determine credit scores and judicial support systems used for parole assessment.
“The proposed regulation is globally unique in its sweeping application to all AI systems, not just systems in specific fields like transport or healthcare,” added Professor Stahl. “It is a clear statement that Europe prioritises its citizens’ fundamental rights over technical autonomy and economic interests.
“There is no doubt this is a bold step in the right direction and is a strong reminder that Europe tries to take seriously its obligation to safeguard its citizens’ fundamental human rights in the light of technological change.”
The SHERPA project,which received €2.8m of funding from the EU, is led by DMU and includes partners from Universiteit Twente (Netherlands), the European Network of Research Ethics Committees (Germany), University of Central Lancashire (Cyprus), Trilateral Research (UK), Stichting Nederlands Normalisatie (Netherlands), Mutual Shoots Ltd (UK), Aequitas Ltd (Cyprus), European Business Summit (Belgium) and F-Secure OYJ (Finland),PineappleJazz (Netherlands).
Posted on Wednesday 21 April 2021