Living and Working with AI News — 2026-09-24 (Thu) Evening News

Overview

Key figures from NVIDIA, Anthropic, and OpenAI spoke about AI risks and safety at meetings hosted by the United Nations Security Council and the British royal family, and the difference in attitude between countries regarding regulatory frameworks (the United States refused regulations, Canada called for international cooperation) became clear. On the other hand, there are parallel reports of the actual introduction of agent-type AI (in the medical field, Alibaba's "agent-type computer") and cases of data leaks and data bullets behind the scenes. From the perspective of corporate management, the focus was on methods for measuring the ROI (return on investment) of AI human resource development and the stagnation in the early stages of AI implementation (the case of the Indian GCC).

Corporate management decision-making

- Alibaba: Announced "agent computer", AI wearable, etc. at 2026 Apsara Conference (product strategy decision making). - Siemens: Announces expansion of strategic partnership with Infrabel (Belgian Railways) to strengthen digital railway operations. - Pabau: Started a service to deploy "AI agent employees" for medical institutions (business decisions related to workforce composition). - Apple: Launched a compensation website for Siri AI delays (management response to product delays).

How the company organization should be

No applicable information

Changes in employee roles

It has been reported that with the introduction of Pabau's "AI Agent Workforce," some of the work of medical institutions will be transferred to AI agents. Klover.ai's analysis also points to ``AI Morale Overload'' in the workplace, or the increased cognitive load associated with gamification (motivation using game elements).

Required competencies

- An editorial in Colorado argues that AI skills should be incorporated into apprenticeships. - China has introduced AI literacy education throughout the school curriculum. - An IBM study found that human-driven AI workflows reduced risk by 18%.

Employee education/human resources development

It is reported that the young graduate market is changing due to AI and new entry routes are needed (International Business Times). Additionally, the Dell-Zinnov report points out that approximately 70% of India's GCCs (Global Capability Centers) are stuck at the AI "pilot stage."

Organizational construction

No applicable information

Division of roles between AI agents and employees

The law firm Ropes & Gray has analyzed the first case of data leakage caused by "agent-based AI" reported in Spain, suggesting issues with the supervisory and auditing system associated with the delegation of authority to AI agents.

Production management using humanoids

No applicable information

ROI concept

- A Reuters editorial says AI could help governments reduce debt, but the effects will be uneven. - At the AI ROI Summit in the contact center field, the general consensus was that ``containment, the idea of completing inquiries using AI and not handing them over to human resources, is over.''

Balance of investment in human resources and humanoids

No relevant information (no heading regarding investment in humanoid robots)

Labor Administration/Worker Protection (Regulatory Authority Perspective)

- British Columbia, Canada, files suit against OpenAI. Regarding the Tumbler Ridge mass shooting, he claimed, ``One phone call would have prevented it.'' - Apple has opened a compensation site for Siri AI delay and provides information on compensation to affected iPhone owners.

Movements in each country

- Canada: Prime Minister Carney asks major AI countries to secure "safe and effective" technology. - United States: President Trump rejects AI regulations and announces the US leadership in the field of superintelligence. - China: Integrating AI literacy education into school curriculum (see above). In the trust survey, the question was ``Do you trust China more than the United States?'' - Australia: Although the government is actively promoting AI, it is reported that the numbers (actual data) show a different reality.

Remarks from VIPs

- Jensen Huang (NVIDIA CEO): If AI companies can't control their models, they will have to "shut down their labs." He also took a pragmatic stance, calling for ``engineering rather than panic.'' It is reported that he was interviewed by the pool before attending a meeting hosted by the British royal family (King Charles). - Dario Amodei (Anthropic CEO): spoke about AI risks at the UN Security Council with Sam Altman (OpenAI CEO). Three founders of EMerge Americas reportedly mentioned AI risks before Amodei. - Sam Altman (OpenAI CEO): Warned at the United Nations that AI could bring about a "renaissance" or "cataclysm."

Integrated analysis: How is it effective for management decisions

The key person's remarks (item 14) are on the topic of ``secure control of AI,'' and Jensen Huang's statement that ``if it is out of control, the laboratory should be shut down'' is directly connected to the Spanish data leak case regarding the delegation of authority to AI agents (item 8). In other words, when entrusting work to an AI agent at the management level, it is suggested that a supervisory and auditing system needs to be developed in advance. On the other hand, IBM's research (item 5) reports that ``human-led workflows'' reduce risk, and when read in conjunction with Pabau's introduction of ``AI agent employees'' (item 4) and the stagnation of the pilot stage in India's GCC (item 6), it can be interpreted that management decision-making is moving toward ``gradual use of AI while leaving human judgment points,'' rather than ``complete delegation to AI.'' The ``end of containment'' in the ROI theory (item 10) also suggests a shift to investment evaluation based on the premise that AI is combined with people, rather than as a complete means of replacing human labor. Regarding regulatory trends (items 12 and 13), while the United States has shown a stance of refusing regulations and international cooperation, Canada is requesting international cooperation, and the need for multinational companies to use different AI introduction policies for each country and region has emerged as a management issue.

Future points of interest

Details of the case of agent-type AI data leakage in Spain and whether or not it affected the development of supervisory systems at other companies. 2. India's GCC exit from AI pilot phase (following Dell-Zinnov report). 3. How will Altman and Amodei's statements at the UN Security Council affect the regulatory policies of each country (including the US policy of refusing regulations)?