The significance of responsible and ethical AI systems has gained immense prominence on the global stage, underscoring the escalating recognition of its far-reaching impact on societies worldwide. Lately, diverse groups and individuals have transitioned from relying solely on Free Software licenses for their projects to pioneering new forms of licensing solutions which impose restrictions related to fields of endeavour, behaviour, community management and commercial practices. This practice has now spilled over to creation of suo moto ethics codes for AI, leading to creation of licenses with restrictive characteristics.
In this talk, I will delve into the concept of openness of software materialized in the historical definition of Free Software and highlight as to how that should not be forgotten while developing an open AI landscape. Furthermore, I implore that licences with restrictive characteristics disable control, transparency, and oversight over technology. This manifests into a negative impact on people’s digital autonomy, distribution of power in society and ultimately democracy. Lastly, I will discuss how the proliferation of the so-called “ethical licenses” poses problems for integration of Free Software, license compatibility and compliance, which affects the whole ecosystem. This presentation seeks to increase awareness about this crucial issue, particularly as public funds are being allocated to AI projects that may, at times, employ restrictive licensing practices.
Additional notes
This talk shall also provide few examples of AI projects calling themselves as open but which are in fact closed. On the other hand, there are AI projects that use free software licenses like Apache 2.0 and are arguably the most maximally ‘open’ AI systems. Such examples shall also be provided.
