When it comes to adoption of artificial intelligence, the US Federal Government is surprisingly advanced. On February 11, 2019, President Trump signed Executive Order 13859 announcing the American AI Initiative, the United States’ national strategy on artificial intelligence. As part of this strategy the US took into consideration principles on Artificial Intelligence published by the Organization for Economic Cooperation and Development (OECD). The AI Today podcast interviewed Adam Murray, International Relations Officer from the US Department of State to discuss these principles in more detail and why it’s important to have AI principles around responsible & trustworthy AI discussed and adopted on an international level.
Adam Murray, US Department of State
OECD AI Principles
Adam Murray is a foreign relations officer and a US diplomat, and has been working with the Department of State for over 13 years. His background includes many postings around the world from Burma to Paris. As the US delegate to CDEP and chair of the OECD Network of Experts on AI (ONE AI) Adam helped craft the OECD principles. For those not familiar with the principles, they focus around five complementary values-based principles for the responsible stewardship of trustworthy AI:
- AI should benefit people and the planet by driving inclusive growth, sustainable development and well-being.
- AI systems should be designed in a way that respects the rule of law, human rights, democratic values and diversity, and they should include appropriate safeguards – for example, enabling human intervention where necessary – to ensure a fair and just society.
- There should be transparency and responsible disclosure around AI systems to ensure that people understand AI-based outcomes and can challenge them.
- AI systems must function in a robust, secure and safe way throughout their life cycles and potential risks should be continually assessed and managed.
- Organisations and individuals developing, deploying or operating AI systems should be held accountable for their proper functioning in line with the above principles.
The recommendations that the OECD created regarding AI are focused around responsible and trustworthy AI. They were adopted this May 2019 and it was really remarkable because it’s the first time a large number of member countries came together to acknowledge their goals for the possibilities and future of AI in the world. While these are guidelines only and not laws that will be heavily enforced, they do seem to carry a lot of ethical weight for those around them, and the peer review process in place will help to keep people and countries on track and implementing them. Adam said at first there weren’t a lot of international voices from governments per se around these principles, but that they had many voices from many backgrounds helping to get these things together. As for member countries, Adam said that many countries were presented including the United States, Singapore, Russia, Dubai, as well as many other voices. Despite the global makeup, the group was able to reach consensus on many different issues.
Digging deeper into the principles
Adam explains the OECD recommendations are divided into two sections. The first section deals with principles for responsible stewardship and trustworthy AI. This section is designed to contribute to the optimistic outlook for how AI can play a big role in the broader wellbeing, economic growth, and innovation in society. It is designed to help promote safety, security and transparency when it comes to using AI in the everyday world.
The second part looks at national policies that governments can implement. These principles look at things like investing in AI research and development, and preparing the workforce for a future where AI is in even broader use. It also talks about how policy can be made and shaped in the future as things around AI continue to develop and change.
As countries move forward with adoption of AI for various processes, decision making, and tasks, questions around transparency and explainability of AI systems are always brought up. What governments can and should do to help push certain policies and regulations around AI forward seems to be a recurring discussion. Adam explains that education and helping to foster a general understanding of how AI can be used should be discussed, so there can be trust built in using these AI systems. Research is currently taking place to make AI more explainable and work from ATARC is moving forward with building transparency assessments for AI models.
Another big effort being made is to find the balance between creating an environment for innovation and creativity while still protecting the civil liberty of Americans. It’s no surprise that data is the heart of AI, so just what data governments use to train AI models should be discussed. The big emphasis that Adam makes is that a bunch of different disciplines be brought in and really take a look at the advancements and changes that are being made and really give perspective to each of them so that they are working the best for all possible stakeholders.
Next steps with OECD
Countries around the world are looking at AI to see how it can help with a variety of different aspects such as advantages to their industrial ecosystem, military and defense, academic institutions, and other strategic areas. As such it’s a priority to get other countries to be open and receptive to American efforts to advance AI. In fact, The United States understands that it must engage internationally to promote a global environment that supports American AI research and innovation.
For the OECD, they’re putting together an observatory so different parties can come in and really see how these principles are being adopted and implemented. Officially launched in early 2020, the OECD Expert Group on AI (AIGO) is made up of members from across the globe. ONE AI is an advisory group providing expert input to the OECD’s analytical work on AI, identifying possible trends and topics around AI. The members provide AI-specific expertise around policy, technical and business topics. The OECD Network of Experts will meet regularly throughout the year and aims to produce deliverables to help push global AI policy forward. As AI continues to become more a part of our daily lives, and governments are finding strategic advantage with the technology, these discussions become even more important.