In recent years, several countries have adopted policies on artificial intelligence (AI) – often called “strategies” – with guidelines on how they intend to regulate algorithm-driven processes, how they will support innovation and education in the AI sector, and how they will address the impact of the design and deployment of AI systems on people’s lives.
When it comes to the development of national AI strategies (NAIS), we believe that safeguards for fundamental rights and freedoms need to be streamlined into their structure. To achieve this, CSOs must be proactively involved in their creation. Civil society involvement brings expertise and examples about the impact of AI-based systems on different groups of citizens, including risks on the ability to exercise their human rights. This contributes to rights-respecting use and deployment of algorithm-driven processes and will support trust in the use of AI-systems throughout societies.
The contents of a NAIS vary from country to country but overall, as explained above, a NAIS is a document that sets out a short, medium or long-term approach to AI and identifies specific areas or priorities that a government will undertake in the following years. Early in 2020, Global Partners Digital (GDP) and Stanford’s Global Digital Policy Incubator (GDPi) published a report that examines governments’ NAIS from a human rights perspective. The report outlines a series of indicators of issues that should be covered in a NAIS and “governance and stakeholder engagement” is one of them. One of the findings of the report is that civil society is usually not actively involved in the strategy development – and when it is, it is not clear how such participation is conducted. This is in contrast with some international and regional standard-setting efforts to incorporate civil society into NAIS.
ECNL and the International Center for Not-for-profit Law (ICNL) have undertaken follow-up research that focuses on the participatory processes in drafting the NAIS in four countries: The Netherlands, Czech Republic, Canada and Australia. These countries were selected as case studies since the above-mentioned GPD/Stanford research points to them as good practices of engagement of civil society organisations in the development of their NAIS.
The purpose of our research is to highlight what extent these countries included civil society during their drafting process as well as in the subsequent implementation and review processes of the NAIS. In particular, we explored the mechanisms, models for participation, perceptions about stakeholders' engagement and the types of civil society organisations that were consulted. This was done primarily through interviews with relevant governmental and non-governmental stakeholders in these countries. Based on our findings and analysis, we recommend good practices for engagement of civil society in further discussions and policy development and implementation of NAIS in these and other countries.
In December 2020 we published our findings, conclusions and recommendations about the participatory processes in two countries, i.e. the Czech Republic and the Netherlands. These countries are member states of the EU and were invited by the European Commission, via the Coordinated Plan on AI in 2018, to prepare their national strategies on the basis of three pillars:
- Boosting technological and industrial capacity on AI;
- Preparing for socio-economic changes brought by AI (which includes AI literacy and relevant education);
- Ensuring an ethical and legal framework in line with the Charter of Fundamental Rights of the EU.
Our research now also includes Australia and Canada. As a next step, we will publish overall recommendations about inclusion, participatory processes and how to embed human rights in NAIS, targeted at all states that develop national strategies on AI or review existing strategies.