Across Europe and Central Asia, governments are accelerating digital transformation and AI deployment. The next step is ensuring these systems align with international human rights standards.
Human rights cannot wait in the context of the progressive adoption of AI in all areas of our individual and social lives. Although human and fundamental rights have been at the core of various legislative initiatives, including the EU AI Act, the effective fulfilment of this commitment still awaits the implementation of concrete instruments to achieve its objective. Setting obligations without providing effective measures does not help public and private entities that want to develop a truly human-centred AI.
In countries where there are no specific provisions on AI and rights protection, and often no legal framework providing full and effective protection of human rights, it is even more difficult to ensure that AI aligns with those rights. Despite numerous declarations, guidelines and strategies, the proposals fall far short of providing concrete solutions to help small and large private and public entities.
For this reason, the UNDP has launched the first international methodology for assessing the impact of AI on human rights. This methodology is the result of work that I started with the UNDP two years ago and has adopted a participatory design approach, based on feedback from experts on the Human Rights Impact Assessment model that I proposed in several regional meetings held in Turkey, Georgia and Bosnia-Herzegovina.
Thanks to this enriching experience, we have been able to design a comprehensive toolkit that covers the entire process, from the Readiness Assessment to the Human Rights Impact Assessment Model. It also includes a Technical Manual to support those lacking specific expertise.
The UNDP Human Rights Impact of AI Assessment Toolkit is not just a document that outlines the key aspects of a potential assessment and the main variables to consider.
The UNDP Human Rights Impact of AI Assessment Toolkit is a ready-to-use toolkit with a step-by-step approach, which makes carrying out the entire assessment process possible, as demonstrated by its successful testing in various countries.
The UNDP Human Rights Impact of AI Assessment Toolkit is not a guidance document, but a toolkit. For this reason, it includes a beta version of an online interface to make the entire methodology accessible to everyone.
Toolkit publication: https://www.undp.org/eurasia/publications/human-rights-impact-ai-assessment-toolkit
Interactive version (beta): https://hria.eu/
