The use of open-source AI provides a number of opportunities and is generally assumed to be drivers of innovation. In particular, the number of jobs in this area has rapidly increased along with research output from industry and the commercialization of that research. It is widely accepted that the changes resulting from these advances will shape society. Diversity in these fields must be increased not only to get people from underrepresented populations into these lucrative jobs but also to have a positive impact by expecting that a more diverse workforce will ensure the fairness of data-driven decisions made by AI and ML algorithms, an issue that has come under scrutiny. Open-source AI also gives rise to a number of challenges that have a tremendous impact on society. This renders an in-depth analysis of the challenges and limits of an open-source approach to AI necessary in order to identify ways ahead.
The aim of the study is to provide an in-depth analysis of AI and open-source solutions facilitating evidence-based decisions and synthesize current state of knowledge to tackle challenges and limitations related to AI and open-source systems. The analysis will discuss the role that open-source could play in accelerating the use and exploitation of AI, giving a critical assessment of the key research and data published on the subject. It will also highlight any point of contention in the public debate, all major stakeholders’ positions and outline policy options.