Artificial Intelligence changing the Future of Policy-Making
/ By Marko Grobelnik, Chief Technical Officer at International Research Centre on Artificial Intelligence, Digital Champion, AI researcher, Jožef Stefan Institute, Mitja Jermol, UNESCO Chair on Open Technologies for OER and Head of Center for Knowledge Transfer, Jožef Stefan Institute, Dobran Božic, Government Office for the Protection of Classified Information and Katja Geršak, Executive Director, Centre for European Perspective (CEP) and Editor-in-Chief of the Bled Strategic Times
This article was originally published as part of the 2021 edition of Bled Strategic Times, the official gazette of the Bled Strategic Forum (BSF) international conference. You can access the full version of this and other BSF publications by visiting our official website.
The unstoppable pace of digitalization, which produces an enormous amount of data that is there to be processed, will have a significant impact on policy-making. Leaders and policy-makers are making decisions that affect our daily lives and our society on every level, from economic life, security, infrastructure, energy, innovation, health, social relations, international relations etc. Policy-making in today‘s fast-paced and interconnected social environment is complex and demands proper decisions and policies fast. The amount of information about the functioning of various aspects of our societies that is now available in digital form and that can be processed is immense. It is far greater than any single person or a group of people can process and analyze using traditional analytical methods. This is where the great leap in Artificial Intelligence (AI) comes in. AI holds the key to improving the process of policy-making.
Challenges for policy-makers
The interconnectedness of the world means that far away events can have an unimaginable impact on our local environment and that events in one area, such as health, will impact all other aspects of our lives, from social and economic to political.
The covid-19 outbreak is a good example that demonstrated immense challenges that the policymakers are facing. The lessons for policy-makers are stark. After 18 months of the global Covid-19 crisis, we can conclude that: the speed of decision making increased enormously, but systems were not prepared for the speed-up. Adaptation to new situations became a necessity, but adaptation procedures were too slow. Solving isolated issues worked well, but without a proper understanding of societal consequences in various other areas. Data to support decision making existed but was not integrated and often not accessible also because of limitations related to legal rights and privacy and security aspects. Technology is available but not used at critical spots.
Policy-making today is in many ways still a process where results are mostly produced manually by individuals, which are more or less knowledgeable about the internal and external environment (but have limited situational awareness), and are mostly oriented to the short-sighted goals (hence also have limited foresight). Therefore, policy-making often feels like correcting mistakes of the past with lots of political compromises in order to achieve desired outcomes. The ‘evidence based’ approach is present but shallow and far from what could be done. The problems that policy-makers are addressing are complex and the ‚manual‘ approach is no longer adequate. This is one of the key reasons why efficient and effective policy-making is hard to imagine without the extensive use of AI technology.
AI leap: solving complex problems
AI is today’s key technology that primarily deals with complex tasks, mainly in the domains where humans can perform, but typically far from optimal. In that sense, AI is taking the role of augmenting human intelligence to the levels which were not in use or even known before. Policy-making of tomorrow is a complex task that combines understanding social, business, political and natural domains, intertwining local, national and global dimensions and the need to react fast with predicted consequences. Without the help of AI technology-driven tools in their daily work, policy-makers will lack situational awareness of such a complex environment that is required to make proper decisions.
Without the help of AI technology-driven tools in their daily work, policy makers will lack situational awareness of such a complex environment that is required to make proper decisions.
The area of science that deals with complex systems is called ‘Complexity Science.’ Although being here for decades already, it is mostly below the radars of public discussions. ‘Complexity Science’ provides strong underlying philosophical and technical fundamentals for studying complex systems be it technical, social, natural or combined. It does not have a commonly agreed definition (as many other scientific disciplines) but resides on some key principles: (a) it takes any kind of human-created innovations without a prejudice, (b) it solves only (or mostly) hard unsolved problems in the nature and society, (c) it is primarily ‘evidence based’ and holistic, creating actionable outcomes.
Joining the current progress of AI with the philosophy and approaches of the ‘Complexity Science’ will likely make the future very ‘explosive’ in terms of the reach and impact of technology. Consequently, the augmentation of human intelligence — the ability to perceive, understand, control and act — will increase significantly. We can safely say the new wave of the so-called ‘holistic AI’ (i.e., in contrast to the ‘narrow AI’ which we mostly know today) will bring about this change in the next 10–20 years.
Today we can spot the early signs of such a future in the most progressive parts of AI like ‘Digital Twins’ and ‘common sense reasoning’ where the goal is to analytically understand the structure of our world beneath the surface and beyond local phenomena. These developments are already fundamentally changing many aspects of humanity thus providing hard challenges for policies and regulations but also opportunities for more informed policy-making. The team at Jozef Stefan Institute (JSI) in Slovenia is deeply and actively involved in addressing these challenges by developing effective AI tools and approaches.
The role of Situational Awareness and Strategic Foresight in policymaking
The aim of the policy-making process is to understand the situation, develop options, solve problems and reach a decision. Therefore, policy-makers need, firstly, to be able to follow events affecting their environment (organization, country, region, world) daily. Secondly, they need to be able to have a birds-eye view of the complex processes and interactions within their environment. Thirdly, they need to be able to understand how external events and what trends might significantly affect their environment. Fourthly, they need to connect all this information into a coherent picture and make decisions for the future of the system. The methods that assist them in doing so are ‘Situational Awareness’ and ‘Strategic Foresight.’
Situational awareness gives us knowledge of what is happening in an organization and around an organization that affects its performance — current and future. Strategic Foresight is the discipline of exploring, anticipating and shaping the future. Policy-makers of today need to complement both methods and use them in their work processes. Policy-makers use of the tools will give them better situational awareness as well as foresight and will have a fundamental impact on the quality of the outcome.
The most groundbreaking leap is the ability to predict the consequences of actions, to predict what will happen or what will be needed in the future. AI technology, predictive analytics, in particular, is already being successfully used in some domains to predict future actions. In the social domain, the team at JSI has already implemented AI to predict the consequences of actions or events. Therefore, if the policy-makers will not use AI technology-driven tools in their daily work they will have neither the situational awareness nor foresight required for proper and timely policy-making.
We already have tools for policymakers today!
An example of an automated ‘policy-making’ tool accessible today is the system produced in the collaboration between OECD and JSI called ‘OECD AI Policy Observatory’ (oecd.ai). It includes a complete and continuously updated global observation of the area of AI across 12 causally interrelated dimensions including academia, patents, projects, companies, investments, job market (supply and demand), education, incidents, media, perception, and policies. Beyond rich visualizations, the tool is getting transformed into the ‘Digital Twin of AI’ capturing influences across the whole ecosystem of the global AI. This will include a projection of the future AI’s impact on society, early detection of innovations that will shape our future, recommendations to the policymakers on how to shape and balance normative side of technology regulation. The Digital Twin is complemented with a kind of a managerial tool. We expect that use of such tools will start growing also in other areas of policy-making in the future using also country-level Digital Twins.
There is a gap between the current technology developments, especially in the AI and related fields, and its usage in policy-making. Policy-makers need to adopt AI into their work processes. By doing so they will have a better understanding of events, trends, and society‘s needs and will be able to ‘play’ different scenarios and select the most proper one. This will give them an advantageous position when planning for the future. There are numerous companies that are already using AI technology to their business advantage. Leaders and decision-makers must keep up with the pace. After all, it is up to them to ensure that our countries and societies prosper. Getting left behind will have drastic consequences for our well-being in the future.