A recent study from the Pew Research Center showed that 53% of people in 20 countries feel that artificial intelligence has been a good thing for society. While over half the world’s population has a positive view of AI, this means that one in every three people in these countries are concerned about the impacts AI can have on society. How do we ensure that AI is trustworthy and its benefits are shared by all? As the statistics show, while there is incremental improvement, there is still a level of hesitancy and suspicion towards AI among the citizens around the world.
In May of 2019 the Organization for Economic Co-operation and Development (OECD) released its Recommendation on Artificial Intelligence (OECD AI Principles). This was the first international standard on AI. It’s purpose is to foster public trust and confidence in AI technologies and realize their potential. One month later, the G20 – which includes key AI players like China and India – welcomed the G20 AI Principles, drawn from this same set of principles.
From then onwards, the OECD set out to help countries implement the AI Principles. It is doing so through the OECD.AI Policy Observatory and its network of experts. OECD.AI is an inclusive hub for public policy on AI that helps countries encourage, nurture and monitor the development and use of trustworthy AI. The OECD.AI network of experts is an international community for discussing common AI policy opportunities and challenges in depth, composed of over 150 experts from around the world.
To gain deeper insights to this important global OECD.AI initiative, I had a deep dive discussion with OECD officials. From the measurement of AI trends and developments to the direction and impact of national and regional AI policies and initiatives, OECD.AI is a prime example of how to move the AI discussion from principles to practice. Its goal is to promote the development and use of AI that is trustworthy and respects human rights and democratic values.
Ensuring that the benefits of AI are shared by all is a mission not exclusive to the OECD. It requires the participation and contributions of AI experts, visionaries and leaders worldwide. Governments and intergovernmental organizations must establish venues for dialogue on AI good practices and use cases such as ITU AI for Good initiative among others. As Vice-Minister, Ministry of Internal Affairs and Communications of Japan, Makiko Yamada stated, “when discussing policies in government, it is essential to be able to cite comparative materials from other countries. The ability to learn about each countries’ AI policies, online, exactly matches the needs of policymakers.”
Worldwide Dialogue and Collaboration on AI
The global nature of AI necessitates global dialogue. Considering that OECD.AI was created in the spirit of collaboration and community, built by experts and practitioners OECD member and partner countries, the EU and numerous other partner countries, and representatives from business, trade unions, civil society and the technical community True to its mission, OECD.AI provides the structure for national governments to act in concert, in a coordinated, coherent manner. Its up-to-date repository of over 600 AI policy initiatives from 60 countries enables the comparison of key elements of national AI policies in an interactive manner.
Mapping AI policies and initiatives by country
At the national level, however, AI strategies often do not reflect a robust understanding of the need for an integrated global approach. For AI to reach its true potential, and for countries across the globe to harness the maximum power of AI there must be cooperation. There must be an open dialogue. There must be open minds among leaders and visionaries to think beyond themselves and to what AI can do for all of humankind.
A stronger emphasis on international co-operation at the national level would contribute to increasing public trust in AI. Though this aspect is starting to be incorporated into national AI strategies, concrete mechanisms need to be put in place to make it happen. In this regard, the OECD.AI network of experts and the “AI Wonk” blog provide a unique space for dialogue.
Developed and developing nations should work in synchrony to ensure trustworthy AI for the benefit of all. This co-operation can be facilitated by OECD.AI’s global reach. Used by policy makers and other stakeholders in over 170 countries, OECD.AI provides evidence-based analysis on AI as a public good. Its AI-related data and metrics, built in partnership with the Józef Stefan Institute in Slovenia, cover over 120 countries. It features national AI policies from Singapore to Argentina to Norway to the EU. This includes AI policies from developing countries in Latin America, Africa, the Middle East and Asia.
Illustration of global data on AI publications by GDP per capita, country, region and over time available on OECD.AI
This global reach is necessary to unlock policy solutions that leave no one behind. In the case of AI policy, inclusiveness is key to mitigating some of the risks posed by AI systems and increasing awareness about the technology’s potential for good.
The impact of OECD.AI is already proving fruitful across the globe. Its resources have been used to help create and shape AI policies and strategies in countries like Canada, Turkey, Colombia and Peru. Its work and indicators have informed and enhanced national and international analysis such as Pan Canadian AI Strategy Impact Assessment, the German AI Observatory, the G20 background paper on Trustworthy AI in Health multiple G20 reports and the recent EC Proposal for AI Regulation.
Armando Guio, CAF Consultant at the Presidency of the Republic of Colombia believes that “the Observatory has rapidly become one of the most important sources of data and knowledge for AI governance.” Melisa Tekeli, Industry and Technology Associate Expert at the Ministry of Industry and Technology, Republic of Turkey says “OECD’s work on AI is the resource I use the most for both my academic studies and work. We think that the visualization and policy comparison tools available on OECD.AI facilitate analysis and save time.” This is what is needed and must continue to expand, countries in Europe and South America both singing the praises of OECD. With this type of common ground and shared purpose the world can only continue to collaborate for the greater good.
One of the working groups of the OECD.AI network of experts is developing practical guidance for AI policy makers by analyzing the EC-OECD database of national AI policies and by leveraging lessons learned and good practices exchanges within the OECD.AI Network of Experts.
The group is focusing on the practical implementation of AI for policy design, policy implementation, policy intelligence, and approaches for international and multi-stakeholder co-operation on AI policy.” Focus and collaboration are two key components to driving real results and effective substantial change. It cannot be stressed enough that only through countries working in concert can the true potential of AI be reached.
Emerging and upcoming initiatives
In addition to all efforts mentioned previously in this article, what is coming next is even more exciting. OECD.AI continues to evolve as a platform for fostering innovation and an open global dialogue. Already, it hosts the OECD’s project on AI in Work, Innovation, Productivity and Skills. This is a wide-ranging programme supported by the German Federal Ministry of Labor and Social Affairs that draws on expertise across the OECD’s technology, labor markets and education domains, informing the international debate on these critical issues.
As I have written previously, digital twin technology has the potential to solve some of the biggest challenges facing humanity. As the number of quality data sources analyzed in the Observatory increases, OECD.AI is seeking to build a “digital twin” of the AI ecosystem to generate predictions and conduct simulations relevant to policy makers. This OECD project will investigate interlinkages and inter-dependencies between different AI-related data streams, along with several official OECD databases, in order to identify correlations and probabilistic causalities to show the true power of AI for the benefit of all.
The OECD.AI network of experts is building a live database of tools to help AI actors rise to the challenge of building and deploying trustworthy AI. These tools vary significantly, from software solutions that detect bias to procedural guidelines for ensuring an AI system is used properly to educational initiatives for AI implementers. This database will be available openly on OECD.AI.
Furthermore, OECD.AI network of experts has launched public consultation to shape its Framework for Classifying AI systems. By the end of June 2021,the goal of the public consultation is to gather all input to improve the framework so it is useful and technically robust for members and its partners.
Other international initiatives have also initiated a strategic call to action in sustainable development, national AI changes and responsible AI. ”Ensuring human-centric AI requires that all of our strengths and mandates are aligned and that we make a complementary effort towards the same end goals. ‘That is why we encourage collaboration with United Nations bodies,” says Karine Perset, administrator of the OECD.AI, AI policy Observatory.
Furthermore, The International Research Centre in Artificial Intelligence under the auspices of UNESCO has launched a call for proposals for listing the Top 100 projects solving problems related to United Nations Sustainable Development Goals with the application of Artificial Intelligence with a focus on Sustainable Development impact.
Action towards Sustainable, Trusted and Responsible AI global development
Every nation that is serious about competing globally in the 21st Century needs recommendations and guidelines to deploy AI in a way that gains acceptance from the public that is trustworthy
Public AI innovation is not an abstract issue that operates in a silo. It determines the ability of our governments to manage crises, to shape the future, to deliver on our needs. The Covid19 pandemic showed that governments matter as does their ability to innovate, especially in AI. Without winning the popular sentiment and developing trust on a national level with the help of OECD.AI, AI may continue to be the subject of controversy and possible extreme legislation that could potentially limit or eradicate the use of AI altogether. During the digital transition around the world, there is a clear need for a fair, independent and balanced education on AI efforts across all OECD members to ensure progress isn’t held back while people’s rights are protected Increasing trust in AI requires a sufficient understanding of the particular challenges posed by the different types of AI systems and the specific contexts in which they operate. The participation of AI experts and the broader public from around the globe is crucial to achieve this understanding.
I believe now more than ever that we would benefit from a more sophisticated, data driven and mature approach to public sector AI innovation to ensure effectiveness of governments and systemically and ensure better outcomes for citizens. The approach OECD.AI is taking is driven by their team and network of trusted AI experts creating and promoting instruments that can be leveraged to facilitate the implementation of the AI Principles. Their goal, trustworthy and responsible AI for the benefit of all, is one that concerns us all.
Moving forward, let’s make our voices heard to make all AI solutions and platforms even more responsible, resilient and trustworthy.