Governance Matters: METI’s report on Governance Innovation states that Japan has been aiming to achieve ‘Society 5.0’. Can you tell us more about what that entails?

Hiroki Habuka: ‘Society 5.0’ is a human-centred society where a high degree of cyberspace and physical space integration can promote economic development and solve societal issues. To achieve this, Japan realises that innovation not only in technology but also in ‘governance’ is essential. With this understanding, Japan recently proposed the ‘agile governance’ framework, which allows  updates to governance mechanisms, such as rules, organisations, and technologies, swiftly with multiple stakeholders.

We live in an era of dramatic change. With the rapid development of digital technology, the quality and quantity of available data and the performance of algorithms has increased dramatically. Digital technology is no longer just a tool but essential infrastructure that substantially affects our behaviour and decision-making.

Cutting-edge technology has the great potential to solve a variety of societal problems, including an aging population, slowing economic growth, widening income disparity, rapid climate change, and the threat of infectious diseases.

However, a variety of risks associated with digital technology have come to light, privacy breeches through the collection and analysis of big data, safety issues due to the autonomous operation of artifical intelligence (AI) systems in  physical spaces, unfair competition because of the concentration of data, threats to fairness due to biased algorithms, and democracy issues as a result of filter bubbles and fake news.

Digital technology is a ‘tool’ — no different to fire or a knife — and depending on how it is used, our society can be utopian or dystopian. When people talk about digital transformation, they are talking about the former, that is, using technology to achieve a better society. To achieve this, it is necessary to maximise the potential of digital technology to solve various societal issues while appropriately managing the associated risks.

Why is it important to manage the risks associated with digital technology and what can be done to manage them?

Imagine a future where tens of thousands of autonomous drones are flying to deliver goods or to help businesses (agriculture, infrastructure maintenance, etc.). To realise this, we need to achieve a certain level of safety.

But what should the rules look like? It is impossible to prescribe detailed rules about the algorithms that each drone can be equipped with. Should the rules pertain to technical standards? If so, the standards cannot be too detailed, and besides, technologies evolve so quickly that the standards might be out of date in a short time.

Or, let’s suppose that two drones collide and fall onto a pedestrian, causing injury. Who should be responsible? The collision might be because of the bad algorithm of either, or both, of the drones. That said, it is also possible that neither of the two drones’ algorithms were wrong, but that the correlation of the algorithms caused an unexpected accident. In that case, who should be punished? The owner, algorithm engineer, data provider, or the manufacturer? But really, does punishing someone really make society better?

This is where innovation in governance comes in. Governance can be defined as “the design and operation of technical, organisational and social systems by stakeholders with the aim of maximising positive impact on society while managing risks at a level acceptable to stakeholders”. Based on this understanding, METI issued a set of reports titled Governance Innovation (GI report). The term ‘governance innovation has three meanings: (1) ‘Governance of Innovation’ in the sense of ‘managing risks’; (2) ‘Governance for Innovation’ in the sense of ‘maximising positive impact’ and (3) ‘Governance by Innovation’ in the sense of using innovative methods for the ‘design and operation of technical, organisational and social systems’. In order to achieve these goals, though, there is a need for fundamental reform of governance structure.

Digital technology itself is a ‘tool’ just like fire or a knife, and depending on how it is used, our society can be utopian or dystopian.

Can you elaborate on why and how METI proposes that these reforms should take place?

Governance is the activity required to realise a variety of ‘goals’ in ‘society’. In the digitalised society, governance has become extremely difficult because both ‘goals’ and ‘society’ are getting more complex, diverse, and fast-changing.

First, we assume that our ‘society’ will change in the way described in the right column of the table below. In short, our society will become more dynamic, uncertain, and global.

Second, the ‘goals’ that we should pursue have become more diversified — the Sustainable Development Goals (SDGs) and Environmental, Social and Governance (ESG) are typical examples. 

Moreover, these ‘goals’ will continue to change in accordance with the development of technology and transformation of society. For example, the meaning of ‘privacy’ has expanded dramatically from a state in which one is not observed by other people, to the right to ask others to treat his/her personal information appropriately.

These characteristics impose difficulties on the traditional governance model, which is based on the belief that “the objectives of governance can be accomplished by regulators who define certain rules in advance, and by citizens who comply with the rules” (a vertical approach).

The world has become too complex to be governed through a single rule or by a single actor. Instead it will be necessary to take a multi-stakeholder approach, where certain goals are shared among stakeholders, and flexible solutions are implemented by each stakeholder (a horizontal approach).

To realise the horizontal governance model, the following two elements will be essential.

• ‍Agile governance cycles: The new governance model needs to keep up with the ever-changing environment, risks, and goals, and; 
• ‍Multistakeholder approach: The new governance model needs to be one in which multiple stakeholders formulate rules and solve problems rather than a single government setting uniform rules.

Can you please explain how the elements of agile governance cycles and multistakeholder approach come together to realise the horizontal governance model?

The GI report proposes the following basic model of agile governance cycles.

‍

This is a model that contains PDCA (Plan-Do- Check-Act) as part of the process (the elliptical cycle in the bottom half, starting from “System Design”). In addition, it requires (i) continuous ‘condition and risk analysis’ and ‘goal-setting’ prior to ‘System design’ (the outer circular cycle), and (ii) sufficient transparency and accountability to external stakeholders (the right bottom line). These requirements are characteristics of an agile governance cycle, which is based on the premise that the environment, risks, and goals are constantly changing, and a multistakeholder approach is necessary to achieve the goals. 

As for the multistakeholder approach, in the traditional governance model, the ‘government’, as the name implies, plays a central role in establishing and enforcing rules. This vertical model is based on the assumption that society is relatively simple and predictable, and that a single entity can prescribe appropriate rules.

However, a society based on digital technology is extremely fast-changing, complex, and uncertain. In such a society, it is difficult for rules to keep up with the speed of changes in technology and business models, and there is a limit to how much monitoring can be done by the government alone. In such a society, a decentralised governance model that emphasises horizontal relationships among stakeholders such as businesses, individuals, and communities is considered necessary.

Under this horizontal governance model, the roles of each stakeholder are expected to change in the following way. 

‍

• ‍The government will serve as a facilitator of multi-stakeholder rule-making, rather than as the sole provider of rules. For monitoring and enforcement, the government will design incentives for businesses, communities, and individuals to proactively take part in those governance processes.
• ‍Businesses will become active designers of rules through providing self-regulations and designing digital architecture, rather than be passive followers of given regulations. They are expected to play a leading role in ensuring trust in new technologies or business models by explaining their rules and architecture externally.
• ‍Communities and individuals are no longer vulnerable actors who lack sufficient information, becoming actors who are able to actively communicate their values and evaluations to society, which empowers democracy.

Are there examples of how the Government of Japan is currently using the Agile Governance framework in shaping the governance of emerging technologies?

Our governance guidelines for Artificial Intelligence (AI) is a good example of how the Agile Governance framework is used by the Government of Japan. The Governance Guidelines for Implementation of AI Principles bridge the gap between AI Principles and corporate practice. As it is necessary to allow for flexible and agile responses to the expected advances in AI technology and changes in social acceptance of AI, the Guidelines provide only basic behavioural goals for implementing AI Principles through agile governance cycles, leaving the specific practices and improvement methods to the companies.

Given that this is still a conceptual approach, what can governments who are interested in Agile Governance do, should they wish to implement it?

It is important to implement agile governance cycles in policy-making or rule-setting. Below, we recommend the following six points that we feel will be particularly important to implementing agile governance in society as a whole:

The ‘rule-based’ approach, in which the law prescribes detailed conduct obligations, cannot govern an increasingly complex digital society. Laws and regulations should take a goal-based approach, which prescribes the goals to be achieved by regulatees. Then, it is the role of each regulatee to determine what specific goals to set and how to achieve them using its technologies and human resources. This is the co-regulation approach.

However, it is not always easy for each company to set its own governance goals and design its own governance systems. Therefore, it is important for public and private sectors to formulate rules jointly with ‘soft law’, such as standards and guidelines. Such soft laws should be used along with goal-based hard laws.

In addition, it is important to actively promote demonstration of experiments using the regulatory sandbox system in order to improve laws and regulations based on practices and evidence.

These laws, regulations, standards, guidelines, etc should be continuously evaluated and improved based on data showing (i) whether they are capable of achieving the policy goals initially set (the inner circle of agile governance cycles), and (ii) whether there is a need to change the policy goals due to changes in social conditions (the outer circle of agile governance cycles).

It is also worth considering reforming corporate sanction systems from the perspective of incentives. A model that focuses only on the outcome of an accident and makes the company responsible will delegitimise corporate innovation, while at the same time creating incentives to conceal information when problems occur. The ultimate purpose of the sanctioning system should not be to identify and punish those responsible, but to prevent future accidents to the maximum extent possible by learning from the ones that have occurred. The sanction system can be designed so that the penalty is imposed based on (i) the quality of governance during ordinary times, and (ii) the extent of cooperation with accident investigations and the implementation of remedial measures when accidents occur.

Finally, disclosure of information is essential for citizens to be able to appropriately reflect their values in governance. Mechanisms to allow individuals to be more substantively involved in political decision-making and governance system design, such as CivicTech platforms, should be implemented.

These are some of many examples of what the government can do in the age of digitalisation. 

After all, ‘governments’ and ‘corporations’ are entities that were invented to achieve better happiness and liberty for human beings. We should use our full imagination to redesign them so that innovation can bring us the maximum positive impact while protecting — and even enhancing — social values. What we need is not a mere update of the existing framework, but dynamic reform of the entire governance framework.