Daily AI Search Memo (2025/11/29 Issue)
Update Date: 2025/11/29
Executive Summary
Tensions are rising over the use of news content as the Japan Newspaper Publishers and Editors Association has strongly demanded that AI operators disclose information regarding the "AI Basic Plan (Draft Outline)" currently being formulated by the Japanese government.
Concrete service offerings and demonstrations incorporating generative AI into business operations are progressing across a wide range of industries, including healthcare, finance, sales, and advertising.
The use of generative AI in the financial sector, such as the strategic partnership between Nomura HD and OpenAI, is in full swing with the goal of creating revenue opportunities.
The use of generative AI in schools is skewed toward a small number of "power user" teachers, revealing a disparity in utilization within schools.
Efforts in education and human resource development are taking concrete shape, such as data/AI talent development programs through university-industry collaboration and hands-on generative AI seminars within universities.
In terms of technology, highly specialized models for specific purposes are emerging, such as Panasonic HD's multimodal AI "LaViDa," models specialized for mathematics, and models specialized for PC operation.
Enterprise solutions compatible with OpenAI's latest model, "GPT-5.1," have been launched, marking the beginning of the implementation phase of the latest models by domestic companies.
Politics Analysis
1. Newspaper Association submits opinion paper on government AI Basic Plan
Source: https://www.okinawatimes.co.jp/articles/-/1722903 (Okinawa Times+ / 2025-11-27)
Key Point: The Japan Newspaper Publishers and Editors Association submitted an opinion paper on November 27 regarding the "AI Basic Plan (Draft Outline)" currently being formulated by the government. They demanded full information disclosure regarding the usage status of news content by AI operators and expressed concern about the decline in reporting functions due to unauthorized use.
Impact: Pressure is mounting on generative AI development companies to disclose the actual usage of news content as training data, which may directly affect future discussions on copyright and licensing frameworks.
2. EU launches "Whistleblower Tool" based on AI Act
Source: https://www.aa.com.tr/en/europe/eu-launches-whistleblower-tool-for-ai-to-ensure-safety-transparency/3752719 (Anadolu Ajansı / 2025-11-27)
Key Point: On November 27, the European Commission launched a "Whistleblower Tool" based on the EU AI Act, allowing anonymous reporting of compliance violations or safety defects in AI systems. Employees of companies and external auditors can now report directly to the AI Office.
Impact: A system is now in place for regulatory authorities to grasp the actual state of AI systems through "insiders" in addition to external monitoring, forcing companies handling high-risk AI to redesign their compliance systems and internal governance. Non-EU companies may also be affected through their bases within the EU.
Economics Analysis
1. Ubie Launches Special Package for 'Ubie Generative AI' for Medical DX
Source: https://prtimes.jp/main/html/rd/p/000000190.000048083.html (Ubie, Inc. / 2025-11-28)
Key Point: Ubie has begun offering a special package for medical institutions that allows them to implement 'Ubie Generative AI' and 'Ubie DPC Supporter' with subsidy application support. The package aligns with medical DX measures included in the FY2025 supplementary budget proposal, supporting the use of subsidies based on hospital functions and regional requirements.
Impact: It may become easier for medical institutions to introduce generative AI and DPC coding support tools, potentially accelerating operational efficiency in medical settings such as medical fee calculation and document creation.
2. Sony Bank Summarizes Inquiry Emails in Under 1 Second Using Generative AI
Source: https://enterprisezine.jp/news/detail/23258 (EnterpriseZine / 2025-11-28)
Key Point: Sony Bank, in collaboration with Sony Group companies and Sony Financial Group, has developed a generative AI application to streamline customer inquiry responses. Using the latest conversational AI model, it summarizes emails in under 1 second per inquiry and has confirmed the ability to generate response drafts for approximately 1,000 inquiries per second.
Impact: The effectiveness of using generative AI in financial institution contact center operations has been demonstrated, which may bring changes to both the cost structure of inquiry handling and the customer experience.
3. Sales Marker Releases Up to 120-Second Video Generation Feature for 'Orcha'
Source: https://saleszine.jp/news/detail/7873 (SalesZine / 2025-11-28)
Key Point: Sales Marker has released 'AI Video' as the third feature of its multi-AI agent 'Orcha,' which automatically generates videos up to 120 seconds long. In addition to existing image and text generation features, it now enables the creation of high-quality video content for sales and marketing.
Impact: Content production in the sales and marketing domain is now automated from text and still images to video, which may support the scale-out of inside sales and ABM initiatives.
4. Nomura Holdings Begins Strategic Partnership with OpenAI
Source: https://jp.reuters.com/markets/global-markets/H67WGYUVJNN27AWLFN2F6KNQG4-2025-11-28/ (Reuters / 2025-11-28)
Key Point: Nomura Holdings announced that it has entered into a strategic partnership with OpenAI, the developer of the conversational generative AI 'ChatGPT.' By combining OpenAI's latest models with Nomura's customer data, the company aims to create new revenue opportunities through applications such as investment advice and financial product design.
Impact: With the partnership between a major securities firm and a foundation model vendor becoming concrete, the competition for generative AI utilization among domestic and international financial institutions may further intensify. Regulatory responses and customer protection frameworks may also become future points of discussion.
5. amana Launches 'AI Creative Architecture' Specialized for Brand Optimization
Source: https://prtimes.jp/main/html/rd/p/000000451.000040283.html (amana inc. / 2025-11-27)
Key Point: Creative production company amana has announced a new service, 'AI Creative Architecture,' which incorporates generative AI into corporate brand management. It integrates four areas—image and video generation, quality evaluation, and operational rule design—to support content production that maintains 'brand identity.'
Impact: By establishing a framework for using generative AI based on corporate brand guidelines, full-scale operation of generative AI in the advertising and promotion fields may advance, potentially leading to a redesign of creative production processes.
Social Analysis
1. Generative AI usage in schools is heavily skewed toward top-tier teachers
Source: https://reseed.resemom.jp/article/2025/11/28/12192.html (ReseEd / 2025-11-28)
Key Point: An analysis of approximately 5 million generative AI usage logs by Study Pocket revealed that generative AI utilization in elementary, junior high, and high schools is concentrated among a small group of teachers. The top 5% of teachers account for about 38% of total usage, and the top 20% account for about 73%, while many teachers hardly use it at all.
Impact: The reality that generative AI utilization in schools is polarized between "specific teacher-dependent" and "organization-wide" models has become clear, highlighting the need to address usage disparities through in-school training and the development of guidelines.
2. Cosmo Energy HD, Databricks, and Shiga University collaborate on AI talent development
Source: https://enterprisezine.jp/news/detail/23255 (EnterpriseZine / 2025-11-28)
Key Point: Cosmo Energy Holdings, data analytics firm Databricks, and Shiga University will launch a joint program in January 2026 aimed at developing data and AI talent in Japan. The program will provide practical education using Databricks' latest curriculum to students in the Graduate School of Data Science at Shiga University.
Impact: By collaborating with energy companies, universities, and tech firms, the construction of a practical, industry-needs-oriented data/AI talent development model is advancing, with expectations for the supply of talent ready for immediate deployment in corporate settings.
3. Shimane University holds a generative AI hands-on seminar
Source: https://www.shimane-u.ac.jp/docs/2025112800018/ (Shimane University / 2025-11-28)
Key Point: On November 27, Shimane University held a "Generative AI Hands-on Seminar" for faculty, staff, and students at the Matsue Campus. Using tools like Microsoft Copilot, they demonstrated automatic email draft generation and image conversion, and also conducted AI application project exercises in small teams.
Impact: By providing a space for practical generative AI experience within the university, it may lead to the formation of concrete images of AI utilization in education and research settings, as well as improvements in the literacy of faculty, staff, and students.
Technology Analysis
1. Panasonic HD and UCLA develop "LaViDa," a diffusion model-based multimodal AI
Source: https://news.panasonic.com/jp/press/jn251127-2 (Panasonic HD / 2025-11-27)
Key Point: Panasonic HD and UCLA researchers have developed "LaViDa," a multimodal AI using diffusion models. It achieves performance exceeding existing autoregressive methods in text generation tasks and high generation efficiency, while also confirming superiority in image understanding and poetry generation.
Impact: The effectiveness of applying "diffusion models" to large language models has been demonstrated, expanding new architectural options in the multimodal field, including text generation. Deployment for practical uses, such as document standardization by in-house AI agents, is also anticipated.
2. JTP's "Third AI" solution now supports GPT-5.1
Source: https://prtimes.jp/main/html/rd/p/000000050.000082223.html (JTP Inc. / 2025-11-28)
Key Point: JTP announced that its "Third AI Generative AI Solution" now supports OpenAI's latest model, "GPT-5.1." It states that by leveraging GPT-5.1's enhanced reasoning and analytical capabilities, the design is flexible enough to handle simple tasks with high-speed processing and complex tasks with persistent reasoning.
Impact: As companies increasingly use multiple models depending on their needs, the growing number of system integrators (SIers) that have already implemented the latest models early on may further accelerate the adoption of generative AI in business assistance and coding support for Japanese companies.
3. Microsoft "Fara-7B": A small-scale model specialized for PC operation
Source: https://indianexpress.com/article/technology/artificial-intelligence/microsoft-unveils-fara-7b-a-lightweight-ai-model-that-can-use-a-pc-from-a-single-screenshot-10390503/ (The Indian Express / 2025-11-28)
Key Point: Microsoft has released "Fara-7B," a small-scale model with 7 billion parameters specialized for PC operation (Computer Use). It is optimized to recognize screens from screenshots and operate GUIs like a human.
Impact: This is a significant step toward realizing "local AI agents" that run on user PCs rather than in the cloud, and it may influence OS and application UI design as well as the future of business automation.
4. DeepSeek "DeepSeek-Math-V2": A math-specialized model with self-verification capabilities
Source: https://indianexpress.com/article/technology/artificial-intelligence/deepseeks-math-v2-ai-model-self-verify-complex-theorems-10390760/ (The Indian Express / 2025-11-27)
Key Point: Chinese AI company DeepSeek has released "DeepSeek-Math-V2," which is specialized for mathematical reasoning. In addition to generating answers, it introduces a mechanism to self-verify proof processes step-by-step, and it reports performance exceeding existing models on high-difficulty mathematical benchmarks.
Impact: In fields requiring strict logic such as mathematics, programming, and law, the effectiveness of a model architecture that "not only provides solutions but also verifies them" has been demonstrated, which may influence the direction of future high-reliability AI design.
Comprehensive Analysis
In terms of politics and regulation, with the opinion paper from the Japan Newspaper Publishers and Editors Association and the launch of the EU's whistleblower tool, the demands for transparency and accountability for generative AI operators are becoming even stronger.
Economically, there is a series of moves toward full-scale adoption of generative AI in diverse industries such as healthcare, finance, sales, and advertising, shifting from one-off proof-of-concepts to a phase of continuous service provision.
Socially and educationally, the disparity in generative AI usage in schools and hands-on seminars at universities has made the gap between "those who use it" and "those who have not yet used it" visible, making training design and human resource development to bridge this gap increasingly important.
Technologically, initiatives toward application-specific models and high-reliability reasoning, such as LaViDa, Fara-7B, DeepSeek-Math-V2, and GPT-5.1-compatible solutions, are prominent, indicating a shift toward diversification away from a sole reliance on massive general-purpose models.
Future Points of Interest
Follow how much of the Japan Newspaper Publishers and Editors Association's opinions will be reflected in the government's "AI Basic Plan (Draft Outline)" and whether the information disclosure obligations for AI operators will be made concrete.
Track how the specific service details, target customers, and risk management designs regarding the strategic partnership between Nomura HD and OpenAI will be disclosed.
Continuously observe how generative AI training and human resource development programs in schools, universities, and companies will lead to correcting usage disparities and practical application in the workplace.
Confirm how new models such as LaViDa, Fara-7B, and DeepSeek-Math-V2 will be incorporated into actual commercial systems and products.