AI governance and human rights

Artificial intelligence (AI) is redefining what it means to be human. Human rights have so far been largely overlooked in the governance of AI – particularly in the UK and the US. This is an error and requires urgent correction. While human rights do not hold all the answers, they ought to be the baseline for AI governance. International human rights law is a crystallization of ethical principles into norms, their meanings and implications well-developed over the last 70 years. These norms command high international consensus, are relatively clear, and can be developed to account for new situations. They offer a well-calibrated method of balancing the rights of the individual against competing rights and interests using tests of necessity and proportionality. Human rights provide processes of governance for business and governments, and an ecosystem for provision of remedy for breaches. The omission of human rights has arisen in part because those with human rights expertise are often not included in AI governance, both in companies and in governments. Various myths about human rights have also contributed to their being overlooked: human rights are wrongly perceived as adding little to ethics; as preventing innovation; as being overly complex, vague, old-fashioned or radical; or as only concerning governments. Companies, governments and civil society are retreading the territory of human rights with a new proliferation of AI ethics principles and compliance assessment methods. As a result, businesses developing or purchasing AI do not know what standards they should meet, and may find it difficult to justify the costs of ethical processes when competitors have no obligation to do the same. Meanwhile, individuals do not know what standards they can expect from AI affecting them and often have no means of complaint. Consequently, many people do not trust AI: they suspect that it may be biased or unfair, that it could be spying on them or manipulating their choices. The human rights to privacy and data protection, equality and non-discrimination are key to the governance of AI, as are human rights’ protection of autonomy and of economic, social and cultural rights in ensuring that AI will benefit everyone. Human rights law imposes not only duties on governments to uphold, but also responsibilities on companies and organizations to comply, as well as requirements for legal remedies and reparation of harms. Companies and investors, governments, international organizations and civil society should take steps to establish human rights as the foundation on which AI governance is built, including through inclusive discussion, championing human rights and establishing standards and processes for implementation of human rights law and remedy in case of breach.