Study on the challenges of open source in the development of European AI

« Challenges and limits of an open source approach to Artificial Intelligence. Brussels, European Parliament Think Tank . »

30-04-2021

Challenges and limits of an open source approach to Artificial Intelligence (europa.eu)

Policy Department for Economic, Scientific and Quality of Life Policies Directorate-General for Internal Policies

Authors: Alexandra THEBEN, Laura GUNDERSON, Laura LÓPEZ-FORÉS, Gianluca MISURACA and Francisco LUPIÁÑEZ-VILLANUEVA

Coupled with the numerous opportunities emerging from the use of artificial intelligence, open source comes with the potential for innovation capacity in both the public and private sector. Advantages include the ability to enhance transparency, facilitate the auditing of AI and thereby enhance citizen trust, while stimulating economic activities and domain-specific expertise. Disadvantages and limits include legal, technical, data, risk management, societal and ethical challenges. This analysis examines all main open source artificial intelligence pro and cons and proposes seven recommendations to boost its uptake.