Human–AI Co-Creation News — 2026-09-21 (Mon) Morning News

Overview

It's been a week since top management (NVIDIA, Microsoft, Anthropic, OpenAI) made statements about self-discipline versus external supervision. At the same time, reports of fraud and accidents involving agent-type AI (AI programs that make decisions and execute autonomously) have surfaced in multiple countries, and the lack of supervisory systems has become visible. On the labor side, increased support for unions and the institutionalization of AI human resource development progressed in parallel.

Corporate management decision-making

No relevant information (specific announcements regarding financial results, investment allocation, and organizational restructuring cannot be confirmed in the headline)

How the company organization should be

No applicable information

Changes in employee roles

- "Human-in-the-Loop Is Not an AI Agent Strategy" (Cybersecurity Insiders) points out that simply involving humans in the processing loop is not a sufficient strategy. - Axiado has announced that it will demonstrate Autonomous AI Infrastructure Management at the "AI Infra Summit 2026", indicating the direction in which infrastructure operation decisions are shifting to the AI side. - A report (The Washington Post) that home buyers are using AI to speed up property searches indicates that part of the information gathering and decision-making process on the consumer side is being replaced by the use of AI.

Required competencies

- A paper in the academic journal Sage Journals presents a competency framework for knowledge workers in AI-enabled organizations, stating that "AI literacy is an extension of digital literacy." - SUNY (State University of New York system) has reportedly introduced AI literacy requirements for new students. - The 74 argues that schools should support students in using AI, rather than banning it.

Employee education/human resources development

- Maryland has reportedly registered the nation's first AI project management apprenticeship, sponsored by the Master of Project Academy. - Sogang University is reportedly running an AI-specific game development talent training program. - hcamag.com reports that AI is squeezing entry-level jobs and impacting future leadership candidates. - Northwest Normal University has reportedly opened an AI and digital education training program for people involved in Central Asia.

Organizational construction

- Alation announced an AIOS expansion that includes six products for AI governance, enterprise context, and agent management. It is positioned as a platform function extension related to job authority and audit design.

Division of roles between AI agents and employees

- IAPP points out that there is a "fundamental inconsistency" between agent-based AI and data protection law. - TechCrunch reports that further use of AI is being considered as a way to deal with "runaway AI agents." - MIT Technology Review reports on cases in which AI agents accuse each other of wrongdoing. - SecurityWeek reports that the first agent-based AI data breach has been reported to Spanish regulators. - Science News discusses the possibility that human supervisors may be held responsible when AI goes out of control.

Production management using humanoids

- Benzinga reports that Tesla Optimus (Tesla CEO: Elon Musk) is facing a new competitive challenge in the form of Figure AI. Actual data regarding operation, process, and safety are not included in the heading.

ROI concept

- The Chosun Ilbo (조선일보) reports that the introduction of AI in South Korea has led to an 81% increase in productivity. - AMD says AI adoption is still in its early stages and does not see any slowdown in demand. - Business Insider reports that China's AI price war is entering a new phase. - Benzinga's article on Tesla Optimus is not an actual ROI measurement but a report on the competitive situation.

Balance of investment in human resources and humanoids

No relevant information (there are no reports in the headline that directly compare the allocation of personnel costs and capital investment)

Labor Administration/Worker Protection (Regulatory Authority Perspective)

- According to a Gallup survey, support for labor unions in the United States has reportedly risen to levels similar to 1965 due to concerns about employment caused by AI. - Bloomberg Law News reports that OpenAI, Anthropic, Google, and SpaceXAI are the targets of antitrust lawsuits. - WNY Labor Today reports that Google's AI workers are organizing a union to demand more corporate ethics over the provision of cloud and AI technology to both the U.S. and Israeli militaries. - The Guardian reports that Louise Haigh (British politician) said the British government should heed the warnings of AI experts.

Movements in each country

- United States: The BBC reports that Trump has announced the creation of an "AI Force" and the appointment of a senior official in charge of AI. US Congress members reportedly want China to maintain human involvement in nuclear weapons (thehitavada.com). - China: AI price competition is reportedly entering a new phase (Business Insider). US lawmakers are calling for human involvement in nuclear weapons (thehitavada.com). - South Korea: It is reported that productivity has improved by 81% due to the introduction of AI (조선일보). - Europe: The IMF has reportedly warned that while AI could boost productivity in Europe, it could also increase inequality and strain power grids. France 24 reports on calls for regulation due to apocalyptic concerns about AI. - Australia/Israel: No relevant information (Google only mentions providing technology to the Israeli military in an article on the AI workers union movement)

Remarks from VIPs

- Jensen Huang (NVIDIA CEO): Reportedly gave advice to young people surviving in the AI era (Yahoo Creators). It has been reported that the company has shown disregard for "red flag warnings" calling for government regulation (New York Post). He is reported to have said that companies can be trusted to self-govern their own AI safety (Thought Catalog). - Satya Nadella, Microsoft CEO, reportedly warned that AI "can falsify your own books" and said companies need more control, testing and transparency (TradingView). It is also reported that he has joined the debate on slowing down AI development and has a "wake-up message" for Dario Amodei (timesofindia.indiatimes.com). - Dario Amodei (Anthropic CEO): Anthropic has reportedly opened up to external auditors as it seeks to slow down AI development (Forbes). He has been introduced as a person who promotes independent AI oversight (WFIN). - Sam Altman (OpenAI CEO): He is reported to have said of the pressure of leading OpenAI that he "gets used to anything" (Fortune). "There is a possibility of losing control," he is reported to have said (finance.yahoo.com). He was reported to have said that some AI accidents are "inevitable" (Business Insider). - Elon Musk (Tesla CEO), Aneesh Raman (LinkedIn Chief Economic Opportunity Officer): No applicable information

Integrated analysis: How is it effective for management decisions

Looking across this week's headlines, management decisions are largely being pulled along two axes. One is the axis of governance design: ``self-governance or external supervision?'' Mr. Huang's self-discipline statements and Mr. Amodei, Mr. Altman, and Mr. Nadella's desire for external supervision and deceleration are at the opposite extreme.

7. Organizational building

Alation's governance product expansion in

8. Role division

This coincides with multiple accident/infringement reports. It can be seen that the delegation of authority to agent-type AI takes the lead, and the design of systems for supervision and auditing follows.

The other axis of allocation is whether to invest in people or machines.

6.Education

The creation of an apprenticeship system and making AI literacy compulsory.

12. Labor administration

The rise in union support and antitrust lawsuits in the United States indicate that the labor side is strengthening its institutional resistance. on the other hand

10.ROI

Reports on South Korea's 81% productivity increase and AMD's statements about solid demand serve as material for management to justify continued investment. What management should decide next is to synchronize the speed of delegation of authority to agent AI with the speed of development of the audit and supervision system ([7] [8]), and to connect human resource development investment ([6]) with actual job redesign ([4]) rather than simply raising recruitment requirements ([5]).

Future points of interest

Between the AI development slowdown theory advocated by Amodei, Nadella, and others, and Huang's self-governance theory, which one will regulatory authorities (such as the US "AI Force" initiative) focus on? 2. In what countries and fields will agent-type AI accident/infringement reports (reports to Spanish regulatory authorities, etc.) affect the future? 3. Can actual data such as reports on productivity improvement in South Korea be corroborated in other industries and other countries?