AI utilization-Beyond AI Research Promotion Organization!

On February 20, the Beyond AI Research Promotion Organization held an online symposium “Living with AI, Going Beyond AI” to commemorate the launch of the Beyond AI Research Promotion Organization.

Beyond AI Research Promotion Organization is an industry-academia research institute established by the University of Tokyo, SoftBank Corp., SoftBank Group Corp., and Yahoo Japan Corporation. have started. Research in two areas: “basic research (medium- to long-term research),” which aims to create new academic fields, and “applied research (high-cycle research),” which aims to apply AI to various social and industrial issues. We aim to contribute to the development of AI research and the realization of a better society.

This time, as a report of the symposium commemorating the launch of the Beyond AI Research Promotion Organization

• panel discussion
• special talk

will be reported.

In this article, we report on a panel discussion attended by Audrey Tang, Minister of Digital Affairs of Taiwan, who has been praised around the world for taking measures against COVID-19 in Taiwan.

*In this article, we have omitted some of the contents of the lecture and changed the expressions.

Three speakers with different backgrounds

In this panel discussion, the discussion ranged from the use of digital technology in the face of the COVID-19 crisis to the topic of “AI ethics,” which has received a lot of attention in recent years.

At the beginning, each of the three speakers introduced themselves about their careers, recent initiatives, and interests.

Mr. Tan emphasized the importance of the concept of “Collective Intelligence” in his work as Minister of State for Digital.

“By trusting citizens and using real-time open data, we were able to disseminate mask sales information. We can democratically solve problems while transforming.This is called assistive collective intelligence.I would like to hear your thoughts on this.”

Sputniko! She talked about her own research theme, “Speculative Design,” along with her own works and issues in contemporary society.

“Starting with raising questions about the future, we discover problems and pose questions, and explore how technology can be useful.”

“Currently, AI is making great progress and development. What I want to ask here today is to make AI free of prejudice, think about how AI can solve many problems, and how it can solve problems while assisting. is.”

Mr. Ikeuchi said that he would like to conduct research at Beyond AI with an awareness of “can we learn more about the neural circuits in the brain?” and “can we make AI development more exciting?”

“I would like to see what people actually want from technology and what they are currently lacking from a social perspective, and then develop technology that fills those gaps.”

Main panel discussion

Bias can already exist during production

First, Mr. Tan talked about fairness, which has become an issue in efforts to prevent the spread of COVID-19 in Taiwan.

Mr. Tan: First of all, I always think about the problem of prejudice (mentioned by Mr. Sputniko!). As a transgender myself, I am subject to such prejudice, so I always think clearly.

The challenge was how to use the Mask Availability Map (*) fairly (as a countermeasure against COVID-19). When I first saw the map, the distribution of pharmacies matched the distribution of the population, so I thought, “This is good.” Only a week later, a member of parliament raised the opinion that they weren’t thinking about public transport times. It looked like the same distance on the map, but the actual cost/opportunity cost was different and it wasn’t fair.

In an evidence-based culture, when biases are found they cannot be hidden, and they may already be there. Sputniko! You also said that earlier.

We will control AI democratically, give feedback, and work constructively to make it public. By doing this, we can ameliorate biases and provide a better offer. We operate a system that allows us to respond to problems within 24 hours. This has played an important role (to prevent the spread of corona infection).

Sufficient feedback is not possible in the current democracy

Sputniko! MR: I think the feedback system is great.

A concern I have at times is whether the commercial incentives will be stronger. Especially when it comes to technology. For example, when you want to create something with fast communication speed, something strong, or something with high efficiency, I think it’s important to be able to give feedback like you said at the same time.

I think we need a system to support the checking of research institutes so that AI can point them out. What do you think? How can we create a sustainable feedback system?

Mr. Tan: Current democracy is designed based on analog systems. It’s based on video and television systems, but these focus on disseminating information and not on feedback. As for feedback, I can only say yes to voting. So the information is very limited.

In order to realize what I think of as “collective intelligence,” it is important for people to participate and give feedback when they find something wrong in the AI ​​society . In Taiwan, half of the population now participates in the feedback system. Even a still-learning 18-year-old can participate without waiting to be old enough to vote, and it helps them understand biases.

Public discussion is necessary to form a society involving citizens

Mr. Tan repeatedly emphasized that it is important for him to send out information himself.

He emphasized the importance of competence, in other words, “to present the agenda to society by oneself”, and said that this was made possible by the feedback system by the Taiwanese people. In response, there was a discussion about what could be done to encourage participation in civil society.

Mr. Hayashi: As an ideal education for children, how can we empower them to participate in civil society?

Mr. Tan: (Omitted) It is important to recognize that algorithms are tools that can change society when citizens use them . Also important, when bias is detected, it must be publicly discussed. We are talking in an academic forum now, but we need these public forums. It has to be done in a way that everyone can see.

Without a forum for discussion like this, I think we would end up in a capitalist society, a surveillance society. It is important to follow the sector with a well-established public infrastructure.

Sputniko! Mr.: I think it is important how to communicate. In recent years, we have also seen extremist ideas online, such as fake news. For example, you are trying to find people who have similar opinions.

Mr. Tan: We make it possible to see the opinions of various people, not just one direction. If there is an agreement, we will proceed in the agreed direction. Since there is a reply button, there is a function of feedback, and trolls cannot exist. We are discussing what the problem is where everyone can see it, so we have a solid understanding of what the ideas are, and we are able to recognize various opinions on the public platform. increase.

It is important to increase interaction with society from the research stage

Hayashi: Professor Ikeuchi, if there is anything from your own research field, please.

Mr. Ikeuchi: There are still few things that can be implemented from our technology. I’m trying to create a scientific solution while thinking about what kind of needs there are. For example, if we could make brain-like cells, they might not have as much bias as the human brain. While developing such possibilities in the future, I would like to discuss and think while receiving opinions. I would like to utilize this research institute to interact with society and share knowledge .

Closing: How can AI integrate a fragmented society?

Mr. Hayashi: This is a question from the three of us. Today, inequality and division are increasing in society. How can AI help increase the sense of solidarity among our citizens?

Tan: The hope of Internet technology is not just being able to download. Listening is not the only advantage. I think it’s wonderful that we can create various programs by ourselves and propose them to the world without obtaining permission.

I think we have to ensure that end-to-end communication that does not require permission is possible.

We have to use AI correctly in that. When we teach young children how to use fire, we teach them how to use it properly, rather than “don’t use it”. It’s the same as that. I think it is important to consider how we can convey our thoughts in order to create a better society.

at the end

There are two points in this panel discussion.

• In Taiwan, actively disclosing information about AI systems and incorporating various feedback while involving the public contributed greatly to preventing the spread of COVID-19.
• In order to involve citizens in such initiatives, it is important to always have public discussions and to have citizens understand that “citizens can change society more conveniently by using AI.”

Audrey Tan’s positive attitude is breathing new life into the digital world. I can’t wait to see what Mr. Tan does next.

In addition, future research at Beyond AI will also be required to create an environment that is conscious of such “dialogue with society.”

Next time, we will report on a special talk between Mr. Teruo Fujii of the University of Tokyo and Mr. Miyagawa of SoftBank Corp.!