"Coexisting with AI" — The era ahead

Behind this shift lies the rapid evolution of AI, transforming the way we interact with information and with each other — from "searching" to "asking AI," and from "working with people" to "working with AI."
We hope that by looking at how AI is influencing various industries and what changes are on the horizon, this will serve as an opportunity to reconsider how we relate to AI.

In November 2022, Open AI launched its conversational AI chat service, "ChatGPT."

This conversational AI chat has captivated people by its ability to read the subtle nuances and unspoken implications within human conversation — something that had previously been out of reach — enabling truly natural dialogue.

The number of users has continued to grow at a remarkable pace, with the number of people using the service at least once a month reaching 100 million just approximately two months after its release.

ChatGPT's release prompted Google to issue a "Code Red" — an emergency alert reserved only for moments when its core business faces a crisis — and in February 2023, Google announced the launch of its own conversational AI service, "Bard," marking the beginning of a major movement in the industry.

Before diving into the main topic, let us take a moment to revisit what AI actually is.
AI (Artificial Intelligence) is a branch of computer science, and refers to the collective term for systems that enable computers to learn and perform tasks normally associated with human intelligence — such as voice recognition, decision-making, and visual perception.
So, when did "Artificial Intelligence (AI)" first come into existence?
Its history is surprisingly long — it is said that the concept dates back as far as the 8th century BC.

The first AI boom took place around the 1950s–1960s in the United States and the United Kingdom, during which AI capable of playing simple games such as mazes and chess was developed. This was followed by a second AI boom in the 1980s, when expert systems gained widespread attention. Today, we are in the midst of a third AI boom. A key driver behind this is the emergence of deep learning — also known as deep neural learning or feature representation learning — through which AI is able to learn on its own. In 2021, the barriers to developing AI models were lowered, and more companies began shifting toward in-house AI development. This has accelerated the "democratization" of AI, further intensifying the demand for efficient creation of training data. Now, let us take a look at how this increasingly democratized AI is impacting our daily lives.

The arrival of ChatGPT has brought sweeping change to the world of education.
Because it can return convincing, natural responses to almost any question, it has made it remarkably easy to produce book reports, essays, and academic papers.
As a result, many concerns have been raised about its impact on learning.
In response, on April 6, 2023, Japan's Ministry of Education, Culture, Sports, Science and Technology announced a policy "to compile reference materials to assist schools in making their own informed decisions regarding the handling of generative AI."

Universities have also been announcing their positions one after another, and how they choose to respond going forward is drawing considerable attention.
Below are a few examples.

◯ The University of Tokyo
Papers and reports produced solely by AI are prohibited.
A forum for discussion on how to use AI and its potential issues will be established.

◯ Tohoku University
"Points to note" have been published separately for students and faculty.
Taking into account that many students will use AI, the university indicates it will take necessary measures and has presented a set of proposed responses.

◯ Sophia University
A "policy on handling AI in academic assessment" has been published.
Unauthorized use in reports, theses, and dissertations is prohibited.
Any confirmed use will be treated strictly as an act of academic misconduct.

What is particularly worth noting here is that none of these institutions have opted for an outright ban on AI.
Rather than prohibiting its use, there is a growing call to rethink how assignments are structured and how learning itself should be approached — with AI use as a given.
In other words, educational institutions are beginning to reshape themselves in order to adapt to AI.
When calculators and the internet first appeared, similar concerns about their impact on education arose.
And yet today, both are used as a matter of course.
It is therefore safe to say that the way we learn will undergo significant change going forward.

So, how exactly will things change?
Personally, I believe attention will increasingly turn to the things — and the qualities — that AI cannot replicate.
While AI excels at generating general answers, it cannot tell us about a person's individual feelings or values.
As a result, perhaps we are moving toward an era where a student's own motivation and the process of learning are valued more highly than test scores.
Ten or twenty years from now, tests may look entirely different — with subjects where AI excels, such as mathematics and science, fading into the background, and fields that only humans can navigate, such as negotiation, taking center stage.

As rising prices — driven by higher raw material and energy costs — continue to bite, more and more companies are making the decision to raise wages.
However, wage increases can have a significant impact on the management of small and medium-sized enterprises.
Some companies, unable to raise wages, are seeing talent leave and are being driven to bankruptcy.
One industry that has long struggled with labour shortages is the taxi industry.
It is also said that, with the spread of autonomous driving, taxi driving is among the jobs most likely to disappear within the next ten years.

Against this backdrop, the taxi industry has begun turning to AI as a way to attract new talent.
The solution adopted by the Tsubame Group is an "AI demand forecasting system."
This system works by training AI on ten years' worth of boarding and alighting data from 1,200 taxis, as well as real-time vehicle availability, and then displaying the required number of taxis and related information on a navigation screen using numbers and colour coding.
Thanks to the introduction of this system, what had previously relied heavily on the experience and intuition of drivers can now provide stable income even for those with no prior experience.
Revenue per driver is reported to have risen by as much as approximately 15%.
The system has also enabled a more flexible approach to working patterns, including options such as 8 to 18 days off per month and day-shift-only arrangements.
Conventionally, more days off means less income — that is simply how things work.
However, the system is said to ensure that a certain minimum monthly income can still be guaranteed.
In other words, the AI system has made genuinely flexible working arrangements possible.
As a result, an increase in job applications from younger workers seeking greater flexibility in how they work is also anticipated.

Company name: Tsubame Anshin Net Group
https://www.tsubame-recruit.com/service/ai_system/

What stands out in this case is that AI has lowered the barrier to entry for highly specialised roles and experience-dependent work, making it easier for people with no prior background to get started.
Lately, it is common to come across articles about "jobs that will disappear in ten years" or "jobs that will be taken over by AI."
However, we believe it is worth shifting our attention toward the jobs that will increase and the work that will become more accessible as AI continues to evolve.
Viewed through that lens, this case study offers a genuinely valuable perspective when thinking about how we might one day "coexist with AI."

In recent years, AI has come to be widely used in medical settings as well.
When it comes to AI adoption in healthcare, the range of applications is vast — spanning diagnosis, treatment, pharmaceutical development, nursing care, and genomic medicine, among others.

Behind the evolution of medical AI lies a growing shortage of personnel in healthcare settings.
Against this backdrop, one development drawing significant attention is the AI doctor.
The UK-based company Babylon Health has announced that its AI doctor sat the MRCGP examination and achieved a score surpassing that of human doctors.
Additionally, on February 9, 2023, a research team from a US medical company announced that when ChatGPT was given questions from the United States Medical Licensing Examination, it achieved a correct answer rate of 52–75%, reaching around the 60% threshold considered the passing mark.
It is expected that performance will improve further as medical knowledge is added and learning methods are refined.

As medical AI continues to make remarkable strides, one concern that naturally arises is whether the jobs of nurses and pharmacists might be displaced.
However, it is generally thought that AI will not completely replace human healthcare workers.
This is because the healthcare environment is an exceptionally sensitive one.
It demands empathy with patients, attentive care, and appropriate communication — all areas where AI falls short.
There are also potential issues that come with being a machine, and if a serious problem were to arise, questions around where responsibility lies would need to be addressed.
For these reasons, at this stage, AI in healthcare is not expected to take over the roles of all medical staff.
That said, there is a growing call for AI to support healthcare professionals, with humans and AI working in tandem to advance the quality of medicine together.

We have looked at how the advance of AI is beginning to affect our lives — but where does the AI development currently underway ultimately lead? We tried putting that question to ChatGPT.

Hearing that answer, we found ourselves imagining a rather unsettling future.
When AI comes to possess intelligence and values that surpass those of humans, there is a possibility that the judiciary, legislature, and executive could come under AI's control.
If that were to happen, the balance of power between humans and AI would surely shift.
If humans came to be dominated by AI, it stands to reason that there would be those who resist.
And eventually, a conflict resembling humans versus AI might break out — ushering in a world not unlike that of The Terminator…
We have let our imagination run rather wild here, but the truth is, a future in which something like that is not entirely out of the question may well be waiting for us.
It is precisely for this reason that we need to start rethinking and updating our sense of ethics now.

Finally, we also asked the question: "Could humans and AGI (Artificial General Intelligence) ever come into conflict?"

Reading the response above, one could interpret it as suggesting that if we fail to remain conscious of ethical and social issues and take appropriate measures, conflict is indeed possible.
Furthermore, if a person with malicious intent were pulling the strings from behind the scenes, it may even be possible to deliberately engineer a state of conflict between humans and AI.
In preparation for such a scenario, it is essential that we do not take information at face value, and instead develop the ability to judge for ourselves what is true and what is not.

We hope this has given you a sense of where things currently stand with the advance of AI.
On April 10, 2023, Sam Altman, CEO of the American venture company "Open AI" — the developer of ChatGPT — attended a meeting of the Liberal Democratic Party's "Headquarters for the Promotion of a Digital Society."
At the meeting, CEO Altman revealed plans to establish a new business base in Japan.
Given this, it is not hard to imagine that the wave of AI adoption will continue to grow in the period ahead.
On the subject of AI, Chief Cabinet Secretary Matsuno emphasised the view that while there are various benefits, new concerns will also arise.
At the same time, the government has indicated a policy of addressing those concerns while also exploring possibilities such as reducing the workload of national civil servants.
With Italy also having moved to temporarily ban the use of ChatGPT, Japan's response to AI going forward will be worth watching closely.
We would be glad if this article has served as an opportunity to recognise the need to think about — and prepare for — how to coexist with AI, as what we once took for granted continues to change at a remarkable pace.

Reference： 『The Much-Talked-About "ChatGPT": Why It's So Impressive — and What Its "Limits" Are』

Reference：『The History of Artificial Intelligence (AI) Research』

Reference：『Google Bard Finally Launches Experimental Version! What Is the Conversational AI That Could Transform Search?』

Reference：『ChatGPT Developer OpenAI CEO Altman Meets with Prime Minister Kishida』

Reference：『Proprietary Models for Every Company』

Reference：『Could "Wage-Hike Bankruptcies" Due to Talent Exodus Be on the Rise? AI-Equipped Taxis Drive Up Revenue and Secure Workforce』

Reference：『Medical AI Already Spans a Wide Range of Applications — How Much of Human Work Will It Come to Handle?』