Why now

Why now — use is spreading, and judgment about how to use it matters

Generative AI is being used more and more, by individuals and by companies. We make no sweeping claims here; we show only primary sources from public bodies.

Adult skills compared internationally — Japan ranks high

PIAAC Cycle 2 (2022-23 fieldwork, published 2024-12-10; 31 countries/economies, ~160,000 adults): Japan ranked 2nd in both literacy (289) and numeracy (291). Across the OECD, only Finland and Denmark showed statistically significant literacy gains; the bottom 10% of performers declined while the top 10% improved, widening within-country inequality

Source
National Institute for Educational Policy Research (NIER), key points of the OECD Survey of Adult Skills (PIAAC), Cycle 2 (OECD 国際成人力調査 第2回調査); OECD press release, 10 December 2024
Survey year
Fieldwork September 2022 – August 2023; published 10 December 2024
Base (denominator)
About 160,000 adults aged 16–65 in 31 countries and economies (nationally representative samples)
Retrieved
2026-09-20

EU AI Act, Article 4 — AI literacy

EU AI Act (Regulation (EU) 2024/1689) Article 4: providers and deployers of AI systems shall take measures to ensure, to their best extent, a sufficient level of AI literacy of staff and other persons operating/using AI systems on their behalf. Per Article 113(a), Chapters I and II (which include Article 4) apply from 2 February 2025. This is EU law and does not automatically apply to companies in Japan.

Source
Regulation (EU) 2024/1689 (Artificial Intelligence Act), Official Journal of the European Union, Articles 4 and 113
Survey year
Adopted 2024; Article 4 applies from 2 February 2025
Base (denominator)
Not applicable (legal text)
Retrieved
2026-09-20

Growth in personal use of generative AI, compared internationally

Individual generative-AI usage experience in Japan: 26.7% (FY2024 survey) to 58.8% (FY2025 survey). FY2024 international comparison: US 68.8% / Germany 59.2% / China 81.2%

Source
Ministry of Internal Affairs and Communications (MIC), White Paper on Information and Communications in Japan, 2026 edition (令和8年版 情報通信白書)MIC, White Paper on Information and Communications in Japan, 2025 edition (令和7年版 情報通信白書; FY2024 survey and international comparison)
Survey year
FY2024 survey (2025 edition) and FY2025 survey (2026 edition)
Base (denominator)
Online survey of individuals by MIC (Japan, United States, Germany, China). Sample size and age composition are not stated in the white paper text
Retrieved
2026-09-20

Japan’s fundamentals are strong. Precisely because of that: even strong fundamentals rust when unused.

What the research says

Concern about new tools is old; the answer is judgment about how to use them

There is no public statistic that shows people "stopped thinking because of AI". What we do know is that there are patterns in how people relate to automation — and that someone has to be able to judge what to delegate.

USE Use Judge what can be delegated, and delegate it
MISUSE Misuse Over-trust without checking
DISUSE Disuse Under-trust; not using what could help
ABUSE Abuse Replace people without considering their role

Figure: four ways of relating to automation (Parasuraman & Riley 1997). The aim is the top left — deciding for yourself when to use.

"Won’t writing weaken memory?" — concern about new tools is 2,400 years old (a classical analogy).

Plato, Phaedrus 274c–275b

A wealth of information creates a poverty of attention.

Simon 1971

The more we automate, the harder the role left to the human becomes.

Bainbridge 1983

People relate to automation in four ways: use, misuse, disuse and abuse.

Parasuraman & Riley 1997

Handing thinking work to a tool (cognitive offloading) can help — when used well.

Risko & Gilbert 2016

The effect of AI is uneven; it depends on the task and on the person’s skill level.

Brynjolfsson, Li & Raymond 2025; Dell’Acqua et al. 2025

Used by tools, or using them

Used by the tool, or using it — in everyday work

We don’t dismiss AI or social media as tools. The difference lies in who decides what gets delegated.

Meeting notes

USED BY THE TOOL Share the automatic summary as-is; nobody checks the decisions.
USING THE TOOL Start from the summary; people confirm decisions and owners, and record them.

Estimates and plans

USED BY THE TOOL Adopt the tool’s numbers as they are.
USING THE TOOL State the assumptions in your own words, and revisit them when they miss.

Explaining to a customer

USED BY THE TOOL Send the generated text unchanged.
USING THE TOOL Adapt it to the customer’s situation and send it, taking responsibility.

Code and documents

USED BY THE TOOL Use what was generated without being able to read it.
USING THE TOOL Read and judge it; decide what to delegate and how to check.

These examples are our suggestions. In training we build them together from your own work situations.

How we can help

How we can help — with our existing services

We do not sell an AI-specific product. Our existing training and coaching support "organizations where everyone can think" and "a culture that masters its tools" (a proposal, not a description of past results).

TRAINING = CLEAR THEORY Customized Training

Using your own work situations, people — including non-IT teams — learn to judge what to delegate, how to check, and how to put assumptions into words. Built from scratch on request.

TRAINING = PUBLIC COURSE Systems Thinking

Learn to see the connections in the whole rather than the parts — for one or a few people.

COACHING = SUPPORT GROUNDED IN REALITY Coaching

We observe real work and give feedback until deciding for yourself becomes a habit.

Key points for your internal brief

Use as a draft for an internal proposal (our suggestion; take figures only from the public-statistics slots).

Premise
Generative AI use is spreading. We do not claim people "stopped thinking". Fundamentals are strong — and rust when unused.
Problem
The criteria for what to delegate to tools are left to each individual.
Approach
Align the criteria through training built on real work situations; make them a habit through coaching.
Target state
Each person can decide what to delegate and how to check.
Evidence
The reference list at the end of this page (author, year) and the public statistics shown on this page (MIC Japan; OECD / NIER; Official Journal of the EU).

Theory behind this page

  • Plato, Phaedrus 274c–275b
  • Simon 1971
  • Bainbridge 1983
  • Parasuraman & Riley 1997
  • Risko & Gilbert 2016
  • Brynjolfsson, Li & Raymond 2025
  • Dell’Acqua et al. 2025

Full list: Theoretical background and sources

We Promote Improvement

The first consultation is free. We listen to your challenges and goals, then propose what fits best.

Get Your Certification

One or a few people? Start with a public course — certified training and Systems Thinking are open for booking.