Saturday, July 20, 2024
HomeAI Talking about the importance of governance and strategic AI utilization!

 Talking about the importance of governance and strategic AI utilization!

In June 2021, the Deloitte Tohmatsu Group established the Deloitte AI Institute (DAII) as an AI research institute. DAII is working with domestic and international AI experts on research activities related to the strategic utilization and governance of AI.

AI utilization is being promoted in all domestic businesses, including the promotion of DX by the Ministry of Economy, Trade and Industry. As AI continues to transform itself, many organizations, including governments, public institutions, and companies, are concerned about how to balance AI and governance (governance and management systems within companies) .

While AI and DX will transform business models and create value for stakeholders, there are other concerns such as bias in decision-making and invasion of privacy.

In a human-centered society, what kind of governance will create ” truly reliable AI “?

This time, we interviewed Mr. Masaya Mori, executive officer of the Deloitte Tohmatsu Group and director of DAII , about the importance of AI utilization and governance .

Table of Contents 

  • DAII with the theme of strategic utilization and governance of AI
    • AI Initiatives Across Businesses of the Entire Group
    • Two themes required for AI utilization
  • Japan lacks awareness of governance
    • Awareness of domestic corporate governance
    • Comparison of Governance with Overseas
    • Introduce an AI framework and strengthen governance
  • Embracing technological evolution and building an ecosystem
    • Technological evolution surpasses human needs
    • Build an ecosystem that incorporates AI
  • Lastly

DAII with the theme of strategic utilization and governance of AI

AI Initiatives Across Businesses of the Entire Group

In addition to consulting services, the Deloitte Tohmatsu Group provides many other services, including audit and assurance services, risk advisory, and financial advisory.

Among them, AI application solutions in areas such as natural language processing, OCR, image recognition , voice recognition , demand forecast/sales forecast, personalization, credit score, fraud detection, accounting audit, loan/screening, M&A support, and asset management. developing.

Before the establishment of DAII, we provided AI solutions within each business, but after the establishment, we began to promote unified AI initiatives across all businesses within the group .

Mori: The use of AI across businesses is progressing. Tasks in a certain area of ​​business are modeled, judged by AI, and executed. If you can standardize tasks, you will be able to standardize them even in businesses in different industries .

Tasks that should be done by humans are done by humans, and tasks that can be handled by AI are left to AI and RPA. I believe that doing so will lead to hyper  automation1, which cross-cuts and automates multiple tasks, and will ultimately lead to corporate transformation.

1Hyper automation: Cross-sectional automation of multiple operations by removing barriers between individual operations by making full use of all kinds of technologies and tools, including AI and RPA.

Cross-departmental corporate transformation through the introduction of AI. It is exactly an initiative that conforms to the essential way of thinking of DX. However, there are very few companies that have achieved this in Japan . The Deloitte Tohmatsu Group provides clients with not only IT, AI/RPA implementation, and data management, but also solutions that incorporate hyper-automation.

The success or failure of introducing AI/RPA depends on the company’s resources. Looking at domestic examples, there are various situations, such as companies that are progressing in the introduction of AI tools as a whole company, companies that are not progressing, and companies that are successful only in specific departments. Under such circumstances, the concept of hyper-automation, which automates the operations of the entire company , becomes important.

Two themes required for AI utilization

ーーMr. Mori, what kind of activities will you be working on at DAII?

Mori: I would like to help companies and society by working on a theme that combines the two DAII advocates of ” strategic utilization of AI ” and ” governance .”

AI is no longer a point-by-point solution, it has the potential to change business processes and value. Therefore, I believe that it is necessary to not only use AI, but also think about how to strategically change the world by making full use of AI .

In recent years, the risks posed by AI have also been discussed. For example, there is talk of who is responsible if automated driving technology causes a traffic accident. In addition, in terms of data bias, there are concerns that AI will be used in a way that lacks racial fairness and encourages discrimination.

Under such circumstances, the strategic use of AI and how to make governance effective will become important themes in the future.

As a foreign-affiliated consulting firm that competes for the top positions in the world, Deloitte has promoted the strategic use and research of AI. In addition, we provide a service called GRC (Governance Risk Compliance) that enhances and streamlines processes such as corporate governance and risk management.

Based on past achievements, we have knowledge on two themes, and we will make use of it to work on research and development of AI.

Japan lacks awareness of governance

Awareness of domestic corporate governance

Japan lags behind other countries in terms of AI governance, but there is a growing awareness in the financial industry that AI is an issue, and there are moves to strengthen governance .

Our clients in the financial industry have traditionally been very conscious of governance and compliance in human operations. Business governance assumes that people do it. When technology such as AI enters here and takes over human work, many clients have a sense of crisis about changes in conventional governance and compliance, so they have come to have an awareness of the problem.

Comparison of Governance with Overseas

ーーWhat is the difference in awareness of governance between Japan and other countries?

Mori: In terms of the use of data and AI, there is a “ prohibition of automatic decision-making ”. This is a topic that is rarely brought up in Japan, but overseas, the GDPR (Personal Information Protection Commission) allows consumers and individuals to refuse automated decision-making decisions.

Compared to Japan, laws and regulations overseas are more advanced than those in Japan, allowing people to choose whether or not to use AI in screening credit cards, mortgages, and university entrance exams .

The insurance industry in the United States and Europe is paying close attention to the impact and risks of automated decision-making, and some companies are self-regulating decisions made by AI.

In terms of the use of facial recognition technology, Japan is being discussed in the opposite direction of automatic decision-making. In Japan, biases arise when people make decisions, so AI makes decisions.

Mr. Mori points out that it is not a big problem that each country has a different way of thinking about AI utilization .

Mr. Mori: The problem I see is that when Japanese companies try to expand globally, they do not understand how AI governance is being done in countries around the world and what is being discussed. and fatal injuries.

Japanese companies need to understand this and advance globally to promote AI research and utilization. Since the network owned by the Deloitte Group has a grasp of governance, we would like to contribute to the growth of our clients’ businesses without neglecting that perspective.

Introduce an AI framework and strengthen governance

Deloitte has released its Global Technology Governance Report 2021 at the World Economic Forum (WEF) 2 . What kind of risks are there in emerging technologies called 4IRT (Fourth Industrial Revolution Technology) such as AI, drones, blockchains, and quantum computers, and what kind of governance is necessary to control them are being discussed. increase.

2 World Economic Forum (WEF): World Economic Forum. An international organization that collaborates with experts in economics, politics, academia, and other areas of society to work to improve the world situation, including economic and environmental issues.

Mori: A framework for preventing unethical use, runaway, and malfunctioning of AI has been discussed and organized, so we would like to further spread it throughout the industry.

Deloitte has also developed an AI framework, and based on that framework, we will introduce governance, organizational movements, and specific solutions to our clients.

Embracing technological evolution and building an ecosystem

Technological evolution surpasses human needs

ーーMr. Mori, why are you working on governance?

Mori: In my previous job, I was in charge of directing a research institute of a major IT company. There were about 70 businesses, and they were also involved in big data and AI solutions.

As I proceeded with my work there, I began to think that the tide had turned. Looking at the trends since the establishment of GDPR, I felt that the disruptive movement of the Internet would require a major shift in the future .

The evolution of technology has gone beyond what people can accept. In the future, when people and technology coexist, I thought that privacy and how to use it for society should be discussed, so I started working on governance that can be involved in that.

Build an ecosystem that incorporates AI

Mr. Mori says that building an ecosystem is important for strengthening governance and promoting the use of AI.

For example, let’s say you have a great sales force with 100% service understanding and 99.9% customer satisfaction. So, let’s say that as a trial, we replace what the sales person is doing with AI, and the customer satisfaction rate is only about 85%. When that happens, the department will not make the decision to replace sales with AI, as customer satisfaction will decline . Many companies stop using AI for this reason.

However, if existing customers can be handled by excellent sales staff, and new customers can be handled automatically by AI, new customers may want to receive even better sales responses.

Then, AI will be able to provide services to 1000, 10,000, and 100,000 new customers with a certain level of quality guaranteed, and excellent sales people will be able to provide close to 100% service.

If you provide services to 10,000 people with an AI customer satisfaction rate of 85%, what those 10,000 people feel will be an important input material for excellent sales. Conversely, by embedding the points when excellent sales people interact with clients into AI, customer satisfaction may improve from 85% to 90%.

The key to successfully running this cycle is to devise ways to combine and interact with conventional business and AI while developing services that make full use of AI to grow the business. . By doing so, the accuracy of AI will continue to improve.

It is difficult for one company to build and turn this cycle alone, and by cooperating with various companies, not only will AI grow, but existing businesses will also improve. Mr. Mori says that it is important to build such a cycle and ecosystem.

In constructing this cycle, ” business standardization ” is a major keyword. By standardizing the know-how of employees who have already achieved 95% or 99% customer satisfaction and applying it to employees with 60% customer satisfaction, it is possible to standardize internal operations at a high level.

Mr. Mori: In building a cycle that strategically combines AI and business, in addition to the hyper-automation I mentioned earlier, the advanced version of Human in the Loop 3 , “Expert Participation AI Utilization,” is effective. I think it ‘s a target.

Until now, big data has been attracting attention, but in reality , I think there are still many know-how and ideas that can only be found in people’s heads .

For example, one of Deloitte’s manufacturing clients is a global electronic components manufacturer. The amazing thing about that company is that it is able to achieve highly accurate demand forecasting and price optimization without using AI at all for demand forecasting. High-quality know-how and high-precision models are in the minds of all sales representatives.

However, one of the challenges faced by companies like this is the inability to get a handle on how to utilize AI where data is not available. In such cases, it is important to verbalize the know-how of experts, annotate small amounts of data, put ideas into logic, and develop models while using them as parameters.

Therefore, an expert participation type AI utilization approach will be required in the future.

3Human -in-the-loop: When machine learning is used, but once labeled data is not recognized, humans intervene to capture exceptional data.


As technology develops, people’s awareness of privacy and governance is changing. It is important to strengthen technology and manage research based on what the public expects from companies and what is being discussed from a global perspective.

Building an ecosystem is also a key to strategically utilizing AI. The quality of the work of all employees can be improved, and those who are already capable can take on the challenge of even more creative work. It can be said that the use of AI to create creative opportunities for employees will lead to corporate growth.


What is an AI algorithm?


Please enter your comment!
Please enter your name here

Recent Posts

Most Popular

Recent Comments