We aim to make AI that actually "thinks" like a human.
Here we introduce the AI research and development currently underway at NTT.
Digital Twin Computing is one of the three major technical fields in the IOWN initiative, along with the All-Photonics Network and the Cognitive Foundation. Here we explain the meaning of Digital Twin Computing and introduce one of its key features, "Human Digital Twins," as well as discussing what Human Digital Twins can be used to achieve in the future.
Looking back over the last 30 years or so, we can see how roughly every 10 years the focus of digitalization alternates between human beings and "things."
So, what will 2025 bring? Following this same pattern, next we should see an advance in "human" digitalization. However, here at NTT we are expecting to see a value explosion as the digitalization of both human beings and things reaches a critical point. We are therefore carrying out research and development based on the concept of "Digital Twin Computing," which we believe will shape that future digital world.
So what is "Digital Twin Computing?" Digital Twin Computing initially starts with the extremely accurate reproduction of real-world human beings and things in cyber space. A virtual being reproduced in cyber space is known as a "Digital Twin." In fact, Digital Twins are already being developed for a variety of applications in different industrial sectors, such as in the automobile sector, for autonomous driving; the robotic control sector; and the medical sector. Digital Twin Computing aims to develop these further and create common processes that allow diverse Digital Twins to be used across sectors.
Digital Twins are electronic data, which means they can be duplicated, merged and exchanged. Digital Twin Computing is the concept of freely mixing various Digital Twins to produce large-scale, highly accurate future predictions that take complex conditions into account. In the future, Digital Twin Computing is expected to make significant contributions to the development of things like autonomous social systems, future urban design, expansions in human capability, and automated decision-making.
Next, we would like to present our work on Human Digital Twins, one of the major features of Digital Twin Computing. With Digital Twin Computing, we aim to create Digital Twins that reproduce not only the outer aspects of a human being, such as physical and physiological characteristics, but also inner qualities such as personality, sensibilities, thoughts and skills. This is because we believe expressing each person's individuality will allow us to elicit interactions based on the diversity that originates from the characteristics of individuals, rather than the interactions between average entities with no individuality. Digital Twins also make it possible to reproduce social aspects in digital space, such as human behavior and communication.
Let's take a look at some examples of what we can do using Digital Twin Computing.
Digital Twins with the personality and characteristics of individual people can react as if they were the real people in response to approaches from others in cyberspace. Conversely, if the Digital Twin is allowed to act autonomously it can also make approaches to others. For example, holding a meeting between Digital Twins can be useful in developing group intelligence, as it can allow you to instantly form an agreement by mixing a diverse range of personalities and expertise without needing to go through people in the real world, and let you do things like visualize trends in the group's thinking in real time.
You can extend the range of human activity from the real world to cyber space using Digital Twin Computing. In cyber space, the Digital Twin behaves as if it were the individual themselves, which makes it possible to create a powerful personal assistant to work on behalf of the real person. The fact that a Digital Twin can be duplicated means it is also possible to work in multiple locations at the same time.
Digital Twins can also be used to communicate with people who do not currently exist, such as the deceased, to gain knowledge and experience. For example, you could communicate with your past and future selves in dialogs that would not be possible in the real world. These dialogs can be used to promote self-understanding, trigger rediscovery of yourself, or acquire ideas, etc., as well as for personal decision-making.
You can also use Digital Twins to interface in cyber space. For example, you could create a derivative of your own Digital Twin endowed with abilities you do not possess (for example, language skills), by exchanging or merging the abilities of your own Digital Twin with someone else's. You can then exploit these newly acquired abilities in the real world, through various devices.
Digital Twin Computing does not consist only of completely new technologies, but also uses various existing technologies such as hardware, software, networks, and AI & IoT. As we carry out our research, we need to take a cross-sectional, big-picture view of a wide variety of technical areas. In addition, when considering the utilization of Human Digital Twins, we also need to scrutinize non-technical aspects in depth—for example, ethical issues, the reliability of simulation outcomes, and the protection of privacy. To achieve this, we will need the cooperation of experts in various fields of research and technology, including social sciences, humanities, natural sciences, applied chemistry, and interdisciplinary fields.
At NTT's Digital Twin Computing Research Center, we are currently conducting R&D on the following three themes: creating Digital Twins with the internal and external aspects of human beings; methods of constructing digital spaces that use Digital Twins; and simulation technology that uses Digital Twins of humans and "things." Moving forward, we will continue to discuss and enhance our understanding of the concept of Digital Twin Computing, and we also intend to work toward promoting the free, reciprocal use of Digital Twins.