AI Highlights — the whole picture — 2026-11-09 (Mon) Daily News

AI stocks are entering a correction phase, and there is growing concern in the US that the boom itself may collapse. In the employment field, reactions to layoffs due to AI are occurring both in terms of systems and discussions, and it was also reported that there is no common yardstick for AI talks between the United States and China. The developments that have been reported in different areas such as stock prices, employment, and governance are converging on a single judgment about how much we can trust AI. The fact that there are multiple reasons for price movements and that there are no uniform methods of measurement are factors that are relevant to this judgment. Please read this half-day article to see what supports your decisions.

Battle and defense over stock price adjustment and employment

In the United States, there is growing concern that the boom itself will collapse due to the correction in AI stocks. The Washington Post reported that Americans are worried about the AI boom collapsing. On the market side, Goldman Sachs sees opportunities in healthcare and non-AI stocks during the correction in AI stocks (The Times of India). Furthermore, the midterm elections caused turmoil in the US stock market, and in addition to AI stocks, alcoholic beverage stocks that were affected by tariffs also became a hot topic (Audacy). Because the reasons for price movements can be divided into multiple factors, including AI, elections, and tariffs, it would be premature to interpret the decline in stock prices as a sign that AI has stalled. Investors will first have to work out which stocks' declines are due to AI and which are due to other factors.

On the other hand, in the employment field, reactions to personnel reductions due to AI are occurring both in systems and discussions. In the Philippines, a senator has introduced a bill that would protect workers from being fired by AI (Philstar.com). In Canada, some argue that employers should not cut jobs due to AI, but that it would be more profitable to reinvest in growth (Toronto Star). The former is based on the idea of restricting layoffs through regulations, while the latter is based on the idea of rejecting layoffs based on management rationality, so their starting points are different. However, both have one thing in common: they do not accept the premise that ``reducing the number of people through AI will increase efficiency.''

To put it simply, stock price adjustments are forcing investors to reconsider their decisions, the bill and reinvestment theory are forcing companies to reconsider their cuts, and the skills report is forcing individuals to reconsider their preparations. Their decisions will depend on how much value and impact AI will actually create. To support this assumption, it is essential to have a system in place, including who governs the functioning of AI and how it is verified. The next section deals with this point.

Lines of governance drawn between countries and authorities

Regarding the US-China AI talks, shattered.io reports that there is only one point of contact and there is no common yardstick for both sides (shattered.io). Even if a forum for consultation has been set up, if there is no consensus on what is considered dangerous and how it will be measured, implementation cannot be confirmed even if an agreement is reached. Two countries, the United States and China, one window, no common yardstick. These three points represent the current state of governance between nations. When a reader encounters the headline ``US-China talks,'' what they should be looking at is not whether the meeting took place, but whether the methods of measurement are being shared.

Domestically, the US Department of Justice has established a new superintelligence task force (ExecutiveGov) to deal with misuse of AI. The creation of a dedicated unit signals a decision to treat abuse as an independent mission, rather than an extension of existing operations. Measures were not shared between countries, and a dedicated organization was established within each country. It can also be seen that the governing body has begun to determine the locus of responsibility within its own country's authorities, without waiting for agreement among nations. The question of who will deal with the misuse of AI is now being answered by a named organization, at least in the United States.

There are similar movements in the private sector and on the debate side. America Magazine asks the question, "Should AI development slow down or accelerate?" (America Magazine) As long as this question remains unresolved, the speed will continue to fluctuate between the judgment of business operators and the design of the system. Meanwhile, Spotify was reported to be labeling AI artists as "AI Persona" (Billboard). This means that one platform has set up its own standards to indicate whether the listener is a human creator or an AI-based entity.

Consultation desks, specialized units, and labeling regulations all require someone to decide in writing ``how much is allowed and where should it be stopped.'' Now that decision-making bodies are divided between the nation, authorities, and companies, whether we can trust AI depends on who sets the standards and who checks them.

Influence on young people, emotions, and bias

Science and Culture Today published an article titled ``AI Chatbots: Dangers Unique to Young People'' about the unique dangers of young people using AI chatbots. The meaning of the distance between AI and people changes depending on the age of the user. The same question extends to the realm of emotions. Bioengineer.org reports that new AI learns emotions from brain waves. The same Bioengineer.org also featured REWAA, a bilingual app that connects mood tracking, AI, and medical examinations. Technology to read emotions and technology to record emotions are being reported side by side. What can be gleaned from these articles is that readers can no longer make judgments based solely on the perspective of the user. It is necessary to separate the risks associated with age and the benefits associated with medical examination.

The relationship between AI and humans has also been reported as a topic of love. Geo News reports that Katie Price has denied rumors that she is dating AI billionaire Felix Henderson. As AI enters our daily lives, even the private lives of those who carry it become the subject of media coverage. This incident is based on a rumor of a relationship and its denial, and does not directly indicate the safety of AI. Therefore, readers need to read the report separately so as not to confuse the reported facts with the evaluation of AI performance.

Design and education are being discussed as ways to stop this impact on people. Britannica covers "What AI designers can do to reduce bias." The Daily Toreador argues that teaching the ethical use of AI is a must in Tech. The former focuses on the design of the creator, and the latter focuses on the education of the user and learner. These two are also potential ways to stop the dangers faced by young people, as reported by Science and Culture Today, and the technology to read emotions, as reported by Bioengineer.org. However, the extent to which it is effective cannot be ascertained from the titles of the articles listed here.

The discussion of design and education to reduce the impact on humans shows that deciding whether to trust AI depends not only on the technology's performance, but also on the mechanisms that support it. In the next section, we will look at how this works.

Accuracy of diagnosis and prediction and spread to the field

The scope of AI predictions is expanding to include Alzheimer's disease, the liver, and agriculture. Neuroscience News reported that an AI model compared to SuperAger predicts Alzheimer's disease up to 15 years in advance. Diagnostic Imaging introduced new research showing that AI analysis of ultrasound test reports is effective in detecting incidental fatty liver disease. This method is used to pick up findings that are different from the original purpose of the test based on the text of the report. Precision Farming Dealer reported that four directions have emerged for the use of AI in agriculture. It is important for those considering the introduction of this technology to keep in mind that predictions are not limited to just one area of medical care, but also extend beyond hospitals. However, to see how accurate it is, it is necessary to read along with the conditions under which it was confirmed.

In research in the cancer field, the method of confirmation is clearly specified for each paper. European radiology reported on virtual liver biopsy to detect steatohepatitis in patients receiving chemotherapy, with comparison with whole liver radiomics and external validation. Cancer Imaging published a study on preoperatively predicting axillary lymph node metastasis using a DCE-MRI-based multimodal fusion model. Gland surgery is involved in the development and validation of a lymph node metastasis prediction model for pancreatic ductal adenocarcinoma that integrates preoperative CT radiomics and clinical variables. Cancers evaluates post-treatment glioblastoma with explainable DSC perfusion radiomics using a soft voting machine learning ensemble. Comparables, external data, and possible explanations allow readers to judge not only the strength of the numbers, but also whether the results hold in other settings.

Images are not the only subject of research. Frontiers in Oncology, a single-cell transcriptome-based prognostic model, identified TM4SF1 as a key factor promoting malignant progression of lung adenocarcinoma. Scientific Reports integrated multi-omics and single-cell transcriptional data to characterize a new subpopulation of senescence-associated and lipid-metabolizing (SALM) macrophages in colorectal cancer. The Journal of Ovarian Research reported a machine learning model that predicts recurrence after fertility-sparing surgery for borderline malignant ovarian tumors. Insights into imaging presents a multicenter study of an MRI-based deep learning radiomics nomogram that distinguishes between FIGO 2023 stage I and stage II-IV endometrial cancer. Multi-institutional studies make it easier to answer the question of whether the results depend on the conditions of a single hospital.

All reports indicate that the scope of application of predictions is expanding, but whether they can be used in the field should be measured by whether they have undergone external verification or multicenter research. As predictions become more widespread, operational arrangements need to be made regarding who will check the results and how they will be held responsible.

The decline in stock prices is due to multiple factors, including AI, elections, and tariffs, and the decline alone cannot be used to determine whether AI has stalled. In employment, there was a reaction to layoffs, and instead of sharing measuring sticks between countries, a dedicated organization was established within each country. It is necessary to read about the dangers to young people, the technology for reading emotions, and the accuracy of diagnoses and predictions, as well as the conditions under which they were tested. What each section has in common is that judgments of trust depend not so much on AI's capabilities or spread, but on how to measure them and whether a system is in place to control them. When receiving numbers or headlines, first confirming the basis and verification conditions will form the basis of your judgment.

Q1: When introducing AI into your business, do you have a procedure in place to check whether the output is correct before implementation? Q2: Are the standards and measurement methods for checking the use of AI shared among the parties involved, and are they left in a form that can be checked later? Q3: When making a business decision based on AI, are you able to explain in your own words the numbers and verification conditions that serve as the basis for that decision?

Finally, three questions

- When introducing AI into your business, do you have a procedure in place to check whether the output is correct before implementation? Q2: Are the standards and measurement methods for checking the use of AI shared among the parties involved, and are they left in a form that can be checked later? Q3: When making a business decision based on AI, are you able to explain in your own words the numbers and verification conditions that serve as the basis for that decision? - Are the standards and measurement methods for checking the use of AI shared among stakeholders and left in a form that can be checked later? Q3: When making a business decision based on AI, are you able to explain in your own words the numbers and verification conditions that serve as the basis for that decision? - When making a business decision based on AI, are you able to explain in your own words the numbers and verification conditions that serve as the basis for that decision?