The Council of Europe released an exhaustive report on the safeguard of human rights with regard to AI. They provide 10 recommendations that are to be implemented by multiple stakeholders to ensure that there are no compromises with human rights as the growth in the field of AI begins to result in real-world. According to the report, although the discussion is aimed towards the member states of the EU, they state that any entity that directly or indirectly has an impact on the development, implementation and effect of AI should consult these recommendations.
The 10 recommendations are as follows:
- human rights impact assessment
- public consultations
- obligations of member states to facilitate the implementation of human rights standards in the private sector
- information transparency
- independent oversight
- non-discrimination and transparency
- data protection and privacy
- freedom of expression, assembly and association, and the right to work
- promotion of “AI literacy”
The specifics of each of these recommendations are fairly exhaustive. For example, the human rights impact assessments are to be conducted by public authorities on systems that they acquire. They even recommend for external audits. Interestingly, with regard to the obligation of member states to facilitate the implementation of human rights standards in the private sector, they recommend reference to the UN Guiding Principles on Business and Human Rights to ensure that there is a non-discriminatory manner of implementation of AI by different actors. These requirements coupled with the requirement of independent regulatory measures and education in the field of AI go a long way in furthering the goals of the EU to maintain accountability and transparency in the instrumentation of AI, while at the same time remaining leaders in the field of AI.
It would be interesting to see if any of these guidelines are implemented in order to regulate AI through a legislative mandate.