The U.S. Announces Endorsement of OECD’s Principles for Responsible AI

Following the announcement of the White House’s Executive Order on Maintaining American Leadership in Artificial Intelligence (EO) and the Department of Defense’s (DOD) Artificial Intelligence Strategy (AI Strategy) in February, as reported on here, the United States recently endorsed the Organization for Economic Co-operation and Development Council’s (OECD) Recommendation on Artificial Intelligence (Recommendation) – the world’s first intergovernmental policy guidelines for Artificial Intelligence (AI). In the Recommendation, the OECD sets forth the “Principles on Artificial Intelligence” to promote innovative and trustworthy AI in harmonization with human rights and democratic values. More than 40 countries have adopted these principles – including all 36 OECD member countries and 6 non-member countries – signaling global cooperation, coordination, and commitment to human ethical and social considerations in promotion of AI. Companies are likely to see more efforts and progress from the White House and around the federal government in support of sustainable, responsible AI.

OECD’s Recommendation – Principles on Artificial Intelligence

With the support of six non-member countries, the OECD hopes that the Recommendation and its five value-based principles will be embraced by any democratic nation, or a nation who shares democratic values, to facilitate an open dialogue on AI. 

The Recommendation identifies five values-based principles for countries to implement in their promotion of reliable AI:

• Inclusive growth, sustainable development, and well-being to benefit people and the planet.
• Human-centered values and fairness that respects the rule of law, human rights, democratic values and diversity, including appropriate safeguards to ensure a fair and just society.
• Transparency and responsible disclosure in AI systems to ensure people understand and challenge AI-based outcomes.
• Robustness, security, and safety in AI systems throughout their life cycles.
• Accountability among organizations and individuals developing, deploying, and operating AI systems.

With these guiding principles, the OECD asks countries to consider the following five recommendations:

• Facilitate public and private investment in research & development to spur innovation in trustworthy AI.
• Foster accessible AI systems with digital infrastructure, technologies, and mechanisms that allow for collaboration with data and knowledge.
• Create an environment to foster the deployment of trustworthy AI systems.
• Empower people with AI skills and support workers in jobs that will employ AI.
• Cooperate across borders and public sectors to ensure responsible control of AI.

In recognizing that countries need assistance in carrying out these principles, the OECD will launch the OECD AI Policy Observatory (Observatory) later this year. The Observatory will be an online live database containing AI resources, from policies and strategies to general information on AI. In addition, countries and other stakeholders will be able to share and update their own AI policies, which will provide an interactive comparison of their respective AI strategies and initiatives. Likewise, the Observatory will provide a platform to discuss and debate AI issues for the international community and other stakeholders.

Significance of the United States’ Support of OECD’s Recommendation

The OECD’s Recommendation is a historic step for the United States and the other member states, and is very significant for the United States as it joins the international community in its pledge for responsible AI. This should come as no surprise, as the Recommendation echoes the White House’s and DOD’s recent announcements on AI. Prior to these announcements, the United States did not have a public position with regard to the ethical and social considerations of AI. But, the United States’ public support of the Recommendation – along with the EO and AI Strategy – demonstrates the United States’ unequivocal commitment to values-based AI. As a result, companies are likely to see more opportunities to partner with the federal government in developing AI. The United States Department of Commerce’s National Institutes of Standards and Technology (NIST) has already taken steps in this direction, and issued a Request for Information for help to create technical standards and tools in consideration of AI technologies. Companies should expect other federal agencies to follow suit in the search for and promotion of responsible AI.