The 127th Installment
The Development of Information Technology and the Future of Society
by Atsushi Shibata,
Assistant Professor
Since there are usually many serious stories about subjects such as research and education, I will honestly write about the future development of information technology.
Recent developments in information technology have been remarkable. It may sound cliché, but the performance of any piece of information technology that we often use is improving day by day. For example, only five years had passed from the start of service for the previous generation 4G communication system in 2015 to the 5G system in 2020. In those five years, the maximum communication speed went from 1 Gbps for 4G to an ideal speed of 100 Gbps for 5G, enabling information to be exchanged 100 times faster. And if we go back 10 years, smartphones only had 1/4 the number of pixels and 1/8 the storage capacity. At that time, communications were at most 10 MBps or less, a difference of 3 digits compared to 4G and 4 compared to 5G. Incidentally, at that time, communication tools such as Line and Slack had not yet appeared, so there were more limitations on what could be done than today. If we look at the past, we may have been able to predict the teleconferencing and various cloud services we have in 2020, but they would have seemed far off in the future.
What will become of this transformation of information technology and society?
Modern society and information technology
Hardware performance improvements such as smartphone storage and CPUs are certainly attractive to engineers, but they are no value on their own. What is important is the services that are provided to users through their performance. For example, even if a smartphone has 100 TB of storage, it will have no impact on our daily lives because there is no way to utilize its performance through current services. In that sense, the artificial intelligence (AI) technology that took off around 2015 has great synergy with modern society, and can be said to have contributed to the emergence of many services.
Deep learning, which has played a role in igniting the boom in AI, is a technology that can achieve high-performance classification and reproduction based on large amounts of data. For example, you can classify what is shown in the image, and reproduce black-and-white photographs into color photographs based on historical image data. Such data, such as text, images, videos, and audio, are scattered across the internet. What this means is that services created using AI have created new data, and based on that data, a cycle has been created in which new services can be achieved through AI.
Of course, AI is not without its problems. There are three issues that current researchers find troubling: fairness, robustness, and explainability.
First, the issue of fairness refers to the problem of unfair classification and prediction due to the data dependence of AI. Classification based on historical data also means that it is not always possible to derive the correct answer for new data. For example, in about 2015, Amazon created a model to predict whether new employees would improve its performance based on past resume data, and tried to use it for recruitment. However, there was a high recruitment rate of men for technical jobs, and as a result, resumes from women had a low acceptance rate. This prevented it from being put to practical use.
The next problem of robustness refers to the resistance against abusing AI to falsify predictions and classification results, or to extract data. This is a major operational challenge, and can encourage crime, such as falsifying facial image recognition to impersonate others, and being able to reproduce the system’s training data.
Last, the problem of explainability refers to the question of whether the results produced by AI can be provided in a way understandable to humans. No matter how good the results of AI may be, they are useless if not understood by the users. If political decisions are left to AI, it will be difficult for people to follow those decisions if they cannot be understood. This also has legal implications. If an autonomous vehicle hits someone, not only would the cause of the accident need to be investigated, but also data to determine who should be held legally responsible.
The one thing that all three issues have in common is the barriers to operating AI in the real world. For this reason, it may be necessary to approach it not only from a technical point of view, but also from a social one, such as how humans will deal with AI.
Considering the future of society and information technology
The field of engineering, not limited to information, has the purpose of making human society better. This means that social conditions have a significant influence on the development of engineering fields. Although mechanical engineering has developed due to the demand for automation, the demand for electricity to power and control it has led to the separation of electrical and electronic engineering, and then to the branching out of information engineering as a new service infrastructure and to process them.
Considering this, in addition to the three previously mentioned problems, it is likely that social changes in line with the new normal and related technologies will develop in the future.
Already, in the past year, data of a different nature is being produced. First, for the analysis of infection pathways, there has been an increase in demand for data such as individual behavior and health conditions. It has been difficult to conduct research in Japan in this field because this data is personal information. In addition, due to telecommuting becoming common, a lot of data is being generated during teleconferences. It is likely that this will facilitate the acquisition of dialog data, of which there has been little in the past, and improve service performance such as for automated response systems. As more services such as home deliveries are accompanied by physical services, services using data linked to individuals may emerge in the future.
Another anticipated problem is the stress caused from being at home. All living things, not only people, are placed under intense stress when their freedom is restricted. Being at home is not only a physical constraint that limits one’s mobility, but also reduces opportunities to meet people and encounter new things. Although mental health issues such as stress tend to be treated as problems for individuals and their families, it seems that there is a demand for services that provide more options for those who are forced to change their lifestyles. In response, there will probably be growth in the field of VR technology, which has long been watched with anticipation. In addition to the size and price of VR equipment, the lack of content is another factor that has been causing its sluggish growth. In contrast, with the current demand from telecommuting, the virtual openness in VR spaces and the direct virtual interactions with others may be quickly converted into content using technology from the field, and lead to an increase in the number of users. In that sense, this may lay the foundation for living in a virtual space such as Second Life, which was popular in the early 2000s.
Furthermore, although the future is uncertain, I feel that the future development of information technology is more about how people interact with it than about technical issues.
It has been 40 years since we started living with information terminals in one hand, and only about 20 years since we have had smartphones. Compliance training has reduced the leakage of personal information, but there are still people who make offensive remarks on social media and those who overly trust information on it. This may be due to the fact that managing information has not yet become commonplace in society as a culture. Although there is demand for technology that places restrictions on information technology, it is not difficult to imagine that the same thing would occur if these restrictions were to be removed for an unforeseen reason. Another reason is that information technology has developed too quickly for people's lives to adjust.
I hope that new ideas and services will emerge that can maintain the relationship between society and information technology in the future.