AI, Symbiosis, Health & Care News — 2026-09-28 (Mon) Evening News

Latest news on co-creation between AI and humans

Latest news on co-creation between AI and humans

Overview

NVIDIA and SAP have successively announced governance and security infrastructure for AI agents, and the focus of corporate implementation has shifted from ``Can it be used?'' to ``Can it be audited?'' It was also a week of notable interactions between top management and politics, with Huang, Musk, Altman, and Cook attending a dinner for the leaders of the United States and China.

Corporate management decision-making

SAP announced that it will work with NVIDIA to address the governance and security of auditable AI agents in enterprise systems through OpenShell. NVIDIA announced the Open Agent Safety Platform, which protects agents from verification to production. MaaseAI has launched a security AI model for enterprise AI governance and protection.

How the company organization should be

No applicable information

Changes in employee roles

There are no details in the headline about what specific operations will change.

Required competencies

An article from Panahoo University in the US discusses AI and ethics from the perspective of ``Don't rewrite this paper for me.'' It has been cited as an issue that tests ethical judgment in academics and writing.

Employee education/human resources development

In Massachusetts, it was reported that the impact of AI could reshape entry-level (new graduates and inexperienced) employment. The urgency of retraining is being discussed for occupations that have been exposed to AI (i.e., a state in which the job content could be replaced by AI). It was also reported that the leaders of OpenAI and Anthropic are scheduled to speak at the UN Security Council regarding AI control risks.

Organizational construction

SAP and NVIDIA announced that they will work to build a governance and security system to ensure the auditability of AI agents in corporate systems. MaaseAI has also introduced a security model for enterprise AI governance. There is no specific design content regarding approval authority or division of responsibility in the heading.

Division of roles between AI agents and employees

MIT Technology Review raised the question, "Who is responsible if an AI agent goes out of control?" A CDW Canada study reports that Canadian organizations are entering a new stage of maturity with the proliferation of agentic AI. It was also reported that cases of AI agents testing security boundaries have been confirmed, strengthening the case for making incident reporting mandatory.

Production management using humanoids

No applicable information

ROI concept

A survey showed that medical institution executives are continuing to invest in technology even as billing for the cost of introducing AI increases. U.S. Treasury Secretary Bessent has reportedly told the Federal Reserve (Federal Reserve) that interest rate hikes can be eased due to productivity improvements brought about by AI. In Türkiye, the AI adoption rate is reported to have exceeded 50%. In either case, it is impossible to tell from the headlines whether the reports are based on actual measured data or estimates.

Balance of investment in human resources and humanoids

No relevant information (no information comparing humanoids or allocation of capital investment and human resources investment in the heading)

Labor Administration/Worker Protection (Regulatory Authority Perspective)

Cases of AI agents testing security boundaries are reportedly increasing the case for mandatory incident reporting (ET Government). MIT Technology Review also raises the issue of who is responsible when AI goes out of control.

Movements in each country

- Canada: CDW research reports that agent-based AI implementation has entered a new stage of maturity. - Türkiye: AI adoption rate is reported to have exceeded 50%. - Australia: The rise of AI-powered shopping agents is reported to be causing caution among major banks. - United States: Impact on entry-level employment in Massachusetts, remarks by the Treasury Secretary, plans for OpenAI/Anthropic leaders to appear at the UN Security Council, etc. were reported.

Remarks from VIPs

- Jensen Huang (NVIDIA CEO): Reported to have attended the US-China summit dinner with Musk, Cook, and Altman. It is also reported that they have issued an ultimatum to OpenAI and Anthropic. - Elon Musk (Tesla CEO): Attended the same dinner. - Sam Altman (OpenAI CEO): Attended the same dinner. - Dario Amodei (Anthropic CEO): It was reported that he had dinner with President Trump at the White House. It is reported that this was achieved despite criticism in the past. - Bill Gates: Reportedly issued a serious warning against accelerating AI development and called for strict regulation.

Integrated analysis: How is it effective for management decisions

Looking at this week's headlines, the issues facing management can be summarized as the need to simultaneously design ``delegation of authority to agents'' and ``location of responsibility.'' Behind the joint announcement of governance and security infrastructure by NVIDIA, SAP, and MaaseAI, there are real concerns about ``Who will be responsible when an AI agent goes out of control'' raised by MIT Technology Review and ``testing security boundaries by agents'' as reported by ET Government. In other words, the more companies introduce agents, the more likely they are to first define the scope of auditability and authority, or risk creating a vacuum of responsibility. In terms of ROI, there is a willingness to continue investing even when AI cost billing becomes serious in the medical field, and statements have been made at the financial authorities to incorporate productivity improvements into policy decisions (pros and cons of easing interest rate hikes), and management investment decisions are beginning to go beyond individual companies and connect with macro policies. The next decisions that management must make are (1) the scope of work to be delegated to agents and its audit system, and (2) the priority of investment in human resources retraining that is commensurate with the expansion of delegation.

Future points of interest

To what extent will NVIDIA/SAP's governance platform for AI agents materialize in actual corporate implementation and audit cases? 2. What the leaders of OpenAI and Anthropic said at the United Nations Security Council and its impact on the discussion of regulations and mandatory incident reporting in each country.

AI and health latest news

Overview

Mental health was limited to promotional announcements of consumer products, with no data showing its effectiveness. In the area of physical health, there were reports questioning the accuracy of calorie counting apps and introductions to laboratory research on aging. In terms of mechanisms, there have been notable moves to introduce systems and workplaces, such as the Japanese Ministry of Health, Labor and Welfare's autonomous AI demonstration test for medical settings, and the announcement of AI implementation policies at an international conference (GDHS 2026).

Mental health

- The portable AI display "NOOKLAB" was announced with the promise of "relieving anxiety about to-do and increasing productivity" (eu.36kr.com). The headline/source is an article introducing the product, and there is no mention of test data showing its anxiety-reducing effect.

Exercise/physical activity

- Market research results have been announced that the "AI in Fitness and Wellness" market will expand at a compound annual growth rate (CAGR) of 20.6% (Market.us). This is a market size forecast, not clinical data showing health benefits. - A new generation version of electronic medical record (EMR) software ``Physitrack'', which is known for remote physical therapy, was announced as equipped with agent-type AI (TradingView). This is a product announcement, and data regarding rehabilitation effects are not listed in the headline.

sleep

No applicable information (The collection source classified the headline ``Bank's risk management officer unable to sleep due to threat from AI agents'' in the sleep category, but the content is a metaphor for financial risk and is not sleep health data, so it has been excluded from this section.)

Nutrition/Meals

- There was a report that an AI app that records calorie intake may underestimate the actual calorie intake by hundreds of kilocalories (Yahoo Creators). It is unclear from the headline which app and conditions were tested, and there is no mention of whether it was a single facility or multiple facilities.

Prevention/Medical checkup

- It was pointed out that even if biomarkers identified by AI appear to be highly accurate, they may fail in larger-scale, practical verification (the "biggest test") (news-medical.net). The target disease, target population, and validation stage cannot be identified from the headline. It is positioned as a research-level cautionary article.

Self-management of chronic diseases

- It was reported that nine students in the Indian state of Tamil Nadu used AI to solve familiar problems, including Parkinson's disease (Samsung Global Newsroom). This is an article introducing a student project, and the method, outcome indicators, and verification stages are unclear from the headline.

Elderly people/nursing care

- Research results were introduced in which AI detected signs of aging from nuclear images of blood stem cells (Medical Xpress/The same content was also distributed as an image release on EurekAlert). This is a laboratory-level image analysis study, and the headline does not mention the relationship with human health indicators or clinical significance. - An article introducing "5 ways" to integrate AI in long-term care (including home care) settings has been published (McKnights Home Care). The article is a practical guide, and there is no data presented to show its effectiveness.

Digital treatment/regulation

- It was reported that the international conference "GDHS 2026" has indicated a policy to focus on AI implementation across the medical field (expresshealthcare.in). There is no mention of the name of the country/authority or specific approval/insurance coverage, just an announcement of the conference theme. - A survey has been reported that medical institution executives are continuing to invest in technology even though they are aware of the increased cost of introducing AI (HealthExec). The size and country covered by the investigation is unclear from the headline.

Workplace health

- It was predicted that AI could become a standard method for safety management in the logistics industry by 2030 (Occupational Health & Safety). This is a predictive article, and there is no data such as actual decline in accident rates in the headline.

Japanese movements

- It was reported that the Ministry of Health, Labor and Welfare plans to conduct a demonstration test in which "autonomous AI" handles handovers between nurses and inpatient adjustments (yomiuri.co.jp). The aim is to address the shortage of manpower in the medical field, and it is in the demonstration stage (prior to trial introduction). - It was reported that Hiratsuka City will introduce AI to create care plans for nursing care insurance and conduct a one-month trial run (wam.go.jp). A trial stage initiative by local governments aimed at streamlining and improving the quality of operations. - Yamashita, a major nursing care equipment rental company, conducted a 2026 Nursing Care AI/Robot Acceptance Survey, and it was reported that 72.5% of caregivers responded positively to the use of AI and robots (Jiji.com). The size and attributes of the survey target are unclear from the headline. - It was announced that N.D. Software, the provider of the nursing care software "Honobono," and BSCode, the provider of the smart AI glass "NAS Glass," have begun collaboration to streamline recording operations (PR TIMES). This is at the stage of announcing the start of the collaboration, and data regarding the effects of its introduction have not yet been presented.

Remarks from VIPs

No applicable information (Although there are references to ``former United Nations cyber negotiator'' and ``Mr. Trump'' in the headline, these are all statements related to cybersecurity, diplomacy, and AI regulation, and we could not confirm that they were statements in the health field.)

AI video/commentary

No applicable information

AI and nursing care latest news

Overview

It was a week of demonstrations and standards development on the system side, with the Ministry of Health, Labor and Welfare moving to demonstrate testing of ``autonomous AI'' that takes over nurses and coordinates hospitalized patients. On the field side, discussions about the use of AI in the field of care management, demonstration experiments that will lead to early diagnosis of dementia, and a survey on acceptance by nursing care product companies were reported in parallel. There were no headlines showing specific introduction and continuation rates for nursing care robots and home monitoring.

AI in nursing care settings

No relevant information (There are no headings that show specific time reduction effects related to records, billing, shifts, and care plans.As a related topic, there is a discussion of care management x AI, but there are no figures for labor reduction in the headings.)

Labor shortage and workers

- It was reported that the Ministry of Health, Labor and Welfare will embark on a demonstration test in which "autonomous AI" will take over tasks for nurses and coordinate inpatient admissions (Yahoo! News). The aim is to address the shortage of medical personnel, but the headline does not provide details on the number of target facilities, the timing of the demonstration, or the scope of the agency. - Regarding PeopleX's "AI interviews," the Ministry of Health, Labor and Welfare announced the compliance status and future policy for the "Business Operations Guidelines" (new standards for AI interviews) scheduled to be applied in September 2020 (PR TIMES). This is a system response to the use of AI in recruiting and securing human resources, and the headline does not mention staffing standards for nursing care workers or direct effects on job turnover. - There are no headings that discuss the division of work that can and cannot be replaced by AI.

Dementia care

- It was reported that a demonstration experiment was conducted for bus drivers that used an app to make a diagnosis in about two minutes, leading to early diagnosis, under the title ``Easy Dementia Determination Using AI'' (Asahi Shimbun). This is a demonstration experiment of "judgment (detection)" and not a "diagnosis" by a doctor. The accuracy of the judgment, the number of people targeted, and the stage of demonstration (prototype/single facility/comparison of multiple facilities) are not specified in the heading. - VRT reports that researchers are supporting the use of AI in dementia care in nursing homes, but this is an overseas (Belgian media) story and will be treated separately from domestic developments (not included in Section 9 of this report, ``Japanese movements''). Detailed effects and number of facilities are not listed in the heading. - There are no specific descriptions in the heading about dealing with BPSD (behavioral/psychological symptoms) or reducing the burden on families.

Home care/monitoring

No relevant information (no headings related to sensors, fall detection, remote monitoring)

Care robots/equipment

No relevant information (There are no headlines indicating the number and continuation rate of transfer support, rehabilitation, and communication robots introduced.As a related matter, Osaka University was reported to have obtained approval for an AI system for early detection of oral cancer, but this is in the field of medical equipment, not nursing care robots/devices, and is not an example of its introduction as a nursing care robot.)

System/Insurance

- The Ministry of Health, Labor and Welfare plans to establish new standards for AI interviews in the "Business Operations Guidelines" effective September 2020, and PeopleX announced the status of compliance with these standards and future policies (PR TIMES). The heading does not say whether the scope of application is the whole or part of the hiring process in the nursing care field. - It was reported that the Ministry of Health, Labor and Welfare will conduct a demonstration test of ``autonomous AI'' that will take over nurses and adjust inpatients (Yahoo! News, overlaps with Section 3). - An article was published (Care Management Online) stating that the use of AI in the care management field is becoming ``full-scale'' and discusses how each field and country should respond to this trend. There is no mention in the headline of the specific system reform or remuneration system. - Yamashita (nursing care equipment rental company) conducted the 2026 Survey on the Acceptance of Nursing Care AI and Robots, and the results showed that 72.5% of people receiving care were willing to use it (Jiji.com). This is the result of a questionnaire survey regarding acceptance intentions, and does not indicate actual implementation or continuation results. The number of survey subjects, method, and definition of ``persons receiving care'' are not listed in the heading. There is no information in the heading regarding consent and privacy.

Japanese movements

- Ministry of Health, Labor and Welfare's demonstration test plan for autonomous AI (nurse handover/inpatient adjustment agency) (Yahoo! News) - Announcement of policy for PeopleX “AI interview” in response to the new standards of the Ministry of Health, Labor and Welfare (applied in September 2020) (PR TIMES) - Essay on how to deal with the situation in the field and the country regarding the full-scale implementation of care management x AI (Care Management Online) - Yamashita's "Survey on the Acceptance of Nursing Care AI/Robots 2026" shows that 72.5% of caregivers are willing to use it (Jiji.com) - Osaka University receives approval for oral cancer early detection AI system (public recruitment guide) *Although it is in the medical device field, not nursing care, it is noted that it is domestically developed. - Demonstration experiment to promote early check-ups for bus drivers using a dementia assessment app (approximately 2 minutes) (Asahi Shimbun)

Remarks from VIPs

No relevant information (there are no quotations with titles in the headings)