Under the Office of the Secretary-General's Envoy on Technology, the new UN High-Level Advisory Body on AI aims to undertake analysis and advance recommendations for the international governance of AI. It strives to be a key multistakeholder body, comprised of professionals from the government, private sector and civil society. To inform the Body, the UN Tech Envoy invited submissions from experts in the industry, governments, civil society and academia, on key global AI governance issues the Body should prioritise.
In our submission, we provided insights based on our experience protecting and promoting civic space – online and offline – and participation of civil society in AI governance.
Our recommendations are as follows:
- Risks must be considered within a rights-based framework, i.e., risks to a broad range of human rights, including political, economic, social and cultural rights, within future efforts to develop AI risk taxonomies and responses. In addition, these efforts must be aligned with the implementation of the SDGs.
- Assessing the impact of AI systems on human rights should be conducted using the already established methods of human rights impact assessment, in line with the UN Guiding Principles on Business and Human Rights (UNGPs). Human rights due diligence, and within those, human rights impact assessment, should be a key component of Global AI Governance, as consistent with the UNGPs.
- To evaluate the risk level of AI systems from a human rights-based approach, several factors must be considered to determine a level of risk. This includes assessing the product design, the severity of the impact, internal due diligence or compliance mechanisms, a causal link between the AI system and an adverse human rights impact, and the potential for remedy, among other factors.
- We urge the UN Tech Envoy and the Advisory Body to prioritise and properly resource stakeholder engagement of civil society and affected communities, both in their own activities and in broader global AI governance, centring at-risk and marginalised groups.
- We recommend that the Advisory Body provide guidelines to outline what specific information AI developers and deployers should report on. These guidelines can inform national regulation by Member-States on transparency provisions of AI development and use, a prerequisite for AI accountability and rights-based global AI governance.
We look forward to following and contributing to the Body’s work, to ensure a human rights-based and inclusive approach to governing AI in a way that centres the needs and risks of marginalised groups.