• Home
  • Services
  • About
  • Contact
  • Blog
  • 知財活動のROICへの貢献
  • 生成AIを活用した知財戦略の策定方法
  • 生成AIとの「壁打ち」で、新たな発明を創出する方法

​
​よろず知財コンサルティングのブログ

開発元非公開のAIモデル「Ox Alpha」

23/8/2026

0 Comments

 
2026年8月21日、開発元非公開のAIモデル「Ox Alpha」がOpenRouterとOpenCodeに突然登場しました。100万トークンコンテキスト・画像/動画入力対応で、1週間無料で、GPT-5.6やClaudeを凌駕するコーディング性能を持つと話題になっており、中国の智譜AI(Z.ai)のGLM系列の未発表のフラッグシップモデルである可能性が高いと噂されています。
この匿名AIモデルについて、生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
“Ox Alpha,” an AI Model from an Undisclosed Developer
On August 21, 2026, an AI model called “Ox Alpha,” whose developer has not been disclosed, suddenly appeared on OpenRouter and OpenCode. The model has attracted considerable attention for its one-million-token context window, support for image and video inputs, free access for one week, and coding performance reportedly surpassing that of GPT-5.6 and Claude. It is widely rumored to be an unreleased flagship model from the GLM family developed by China’s Zhipu AI (Z.ai).
I asked generative AI to conduct an in-depth investigation into this anonymous AI model. Please note that the findings and analysis are based solely on publicly available information and may not necessarily reflect the actual situation. They may also contain inaccuracies.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

生成 AI に基づく進歩性判断のモデル化評価

22/8/2026

0 Comments

 
月刊「パテント」2026年6月号の「生成 AI に基づく進歩性判断のモデル化評価」は、生成AIを活用して特許の「進歩性」を判断する手法を検証した実験報告です。著者は、法的判断の思考手順を言語化したフレームワークとしてプロンプトに実装し、AIがどの程度一貫性のある結論を出せるかを考察しています。実験の結果、AIは構成要素の対比や差分の抽出において高い効率性を示す一方で、図面の理解や技術常識の適用には人間による補完や対話が不可欠であると判明しました。また、客観的な資料に基づく判断プロセスを構築することで、出願戦略の最適化や実務作業の負担軽減に繋がる可能性が示唆されています。
この報告を生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
生成 AI に基づく進歩性判断のモデル化評価
https://jpaa-patent.info/patent/viewPdf/4836
 
 
Evaluation of a Generative AI–Based Model for Assessing Inventive Step
“Evaluation of a Generative AI–Based Model for Assessing Inventive Step,” published in the June 2026 issue of the monthly journal Patent, is an experimental study examining a method for using generative AI to assess the “inventive step” of patent applications. The authors translated the reasoning process involved in legal determinations into an explicit framework, incorporated it into prompts, and examined the extent to which AI could produce consistent conclusions.
The experiments showed that AI can perform highly efficiently in comparing claim elements and identifying differences. At the same time, they revealed that human input and interactive dialogue remain essential when interpreting drawings and applying common general technical knowledge. The study also suggests that establishing a decision-making process grounded in objective evidence could help optimize patent filing strategies and reduce the workload involved in patent practice.
I asked generative AI to conduct an in-depth analysis of this report, and the results are presented for your reference. Please note, however, that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect the actual circumstances, and may contain inaccuracies.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

防衛デュアルユース技術の知財戦略

22/8/2026

0 Comments

 
防衛・民生両用(デュアルユース)技術を保有する企業が、競争力維持と安全保障を両立させるための知的財産戦略は、特許の出願非公開制度や政府調達における技術データ権など、権利の「所有」ではなく「利用権の設計」を重視することが必要です。
防衛デュアルユース分野における知財戦略について生成AIに深掘りさせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
Intellectual Property Strategies in the Defense Dual-Use Sector
For companies that possess dual-use technologies applicable to both defense and civilian purposes, an intellectual property strategy that balances the maintenance of competitiveness with national security must emphasize not simply the “ownership” of rights, but the careful structuring of “rights of use.” Key considerations include the non-disclosure system for patent applications and rights in technical data under government procurement contracts.
I asked generative AI to conduct an in-depth analysis of intellectual property strategies in the defense dual-use sector. Please refer to the resulting report. Please note, however, that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect actual circumstances, and may contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

防衛デュアルユース

21/8/2026

0 Comments

 
NIKKEI Tech Foresightの連載「防衛デュアルユース」(全5回)では、防衛投資とデュアルユース技術が先端技術と産業へ及ぼす影響についての予測が解説されていました。
【第1回】防衛デュアルユース、戦略17分野と密接 5つの注目技術
【第2回】米国のデュアルユース、2040年に500兆円規模 実態と今後
【第3回】AIが防衛作戦、代表格パランティアは何者か 民生展開も
【第4回】防衛技術をフィジカルAIへ生かす 米アンドゥリルに学ぶ
【第5回】日本の防衛デュアルユース、有望3分野 R&D再考で産業へ
「防衛デュアルユース」について、生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
防衛デュアルユース、戦略17分野と密接 5つの注目技術
https://www.nikkei.com/prime/tech-foresight/article/DGXZQOUC311XP0R30C26A7000000
                       
 
Defense Dual-Use Technologies
The five-part NIKKEI Tech Foresight series titled “Defense Dual-Use Technologies” examined how defense investment and dual-use technologies are expected to affect advanced technology development and industry.
Part 1: Defense Dual-Use Technologies Closely Linked to 17 Strategic Fields: Five Technologies to Watch
Part 2: The U.S. Dual-Use Market Could Reach ¥500 Trillion by 2040: Current Landscape and Future Outlook
Part 3: AI in Defense Operations: What Is Palantir, the Leading Player, and How Is Its Technology Expanding into Civilian Markets?
Part 4: Applying Defense Technologies to Physical AI: Lessons from U.S.-based Anduril
Part 5: Japan’s Defense Dual-Use Sector: Three Promising Areas and the Need to Rethink R&D for Industrial Growth
I asked generative AI to conduct an in-depth analysis of defense dual-use technologies, and I invite you to review the results. Please note, however, that the research and analysis conducted by generative AI are based solely on publicly available information. They may not necessarily reflect actual circumstances and may contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

Claudeがタンパク質設計と分析化学を加速

21/8/2026

0 Comments

 
2026年8月18日、AnthropicはフロンティアAIモデルである「Claude Mythos Preview」および「Claude Opus 4.8 / Opus 5」が、生命科学分野における高度な研究課題を自律的に遂行したとする最新の研究成果を公開しました。本発表は、医薬品開発の最初期段階に位置する「タンパク質バインダーの de novo(ゼロからの)設計」と、日常的な研究業務の多くを占める「分析化学データの解釈・処理」の2領域において、AIエージェントが人間の専門家に匹敵あるいはそれを凌駕する精度と効率を発揮したことを実証しているとしています。
この件について、生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
How Claude is accelerating protein design and analytical chemistry
https://www.anthropic.com/research/Claude-accelerates-protein-design
 
 
Claude Is Accelerating Protein Design and Analytical Chemistry
On August 18, 2026, Anthropic released its latest research findings, reporting that its frontier AI models—Claude Mythos Preview and Claude Opus 4.8 / Opus 5—had autonomously carried out advanced research tasks in the life sciences. According to Anthropic, the results demonstrate that AI agents achieved accuracy and efficiency comparable to, or even surpassing, those of human experts in two areas: the de novo design of protein binders from scratch, which represents one of the earliest stages of drug discovery, and the interpretation and processing of analytical chemistry data, which account for a substantial portion of routine research work.
I asked generative AI to investigate this topic in greater depth, and the resulting analysis is provided for your reference. Please note that the research and analysis generated by AI are based solely on publicly available information and may not necessarily reflect the actual circumstances. They may also contain inaccurate information.
 
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

現在のAIエージェントは自律的オープンエンド研究を遂行できない

21/8/2026

0 Comments

 
「Can AI agents conduct open-ended AI research?」という論文は、AIエージェントが自律的にオープンエンドな研究を遂行できるかを検証した、プリンストン大学などの研究チームによる調査報告です。著者らは、未発表の高品質な論文の問いをAIに解かせ、その成果を原著者が査読する「シャドウ・エボリューション(影の評価)」という新たな評価手法を提唱しており、実験の結果、最新のAIは実験環境の構築といった高度なエンジニアリングには長けているものの、研究の核心である独創的な問いへの進展や論文としての質においては、専門家から「明確な拒絶」を受ける水準にとどまることが示されました。
AIによる研究の自動化には、現時点のモデルでは克服できない重大な限界があるとしています。
この論文を生成AIに深掘りさせましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
[Submitted on 29 Jul 2026 (v1), last revised 7 Aug 2026 (this version, v2)]
Can AI agents conduct open-ended AI research? Early evidence from two case studies
https://arxiv.org/abs/2607.27191
 
 
Current AI Agents Cannot Autonomously Conduct Open-Ended Research
The paper “Can AI Agents Conduct Open-Ended AI Research?” reports on a study conducted by a research team from Princeton University and other institutions to examine whether AI agents can autonomously carry out open-ended research. The authors propose a new evaluation method called “shadow evaluation,” in which AI systems are tasked with addressing the research questions of high-quality, unpublished papers, and their results are then reviewed by the original authors.
The experiments showed that although the latest AI systems are highly capable of performing sophisticated engineering tasks, such as setting up experimental environments, they remain unable to make meaningful progress on the original research questions at the heart of the studies or produce work of publishable quality. Expert reviewers judged the AI-generated research to be at a level warranting unequivocal rejection.
The authors conclude that the automation of research using AI still faces fundamental limitations that cannot be overcome by current models.
I asked generative AI to conduct an in-depth analysis of this paper, and I hope you will find the results informative. Please note, however, that the research and analysis generated by AI are based solely on publicly available information. They may not necessarily reflect the full reality of the subject and may contain errors.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

NTTデータのAIを活用した発明発掘支援サービス

20/8/2026

0 Comments

 
NTTデータは、企業の知的創造サイクルの高度化に向け、知財業務の変革を支援するサービスの提供を2026年10月以降に開始することを公表しました。このサービスは、研究開発部門と知財部門をつなぎ、発明相談から知財活用までの業務を支援するもので、第一弾として、AIを用いて発明相談から発明提案書作成までのプロセスを円滑にする発明発掘を支援するサービスを提供するということです。
生成AIに本サービスについて深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
AIを活用した発明発掘を支援するサービスを先行提供開始
~研究開発部門と知財部門をつなぎ、企業の知財業務変革を支援~
https://www.nttdata.com/global/ja/news/topics/2026/081900/
 
NTTデータ、AIで企業の知財業務を支援 発明の発掘を効率化
2026年8月19日
https://www.nikkei.com/article/DGXZQOMG00015_Z10C26A8000000/?utm_source=dlvr.it&utm_medium=mastodon
 
 
NTT DATA’s AI-Powered Invention Discovery Support Service
NTT DATA has announced that, beginning in October 2026 or later, it will launch services designed to support the transformation of corporate intellectual property operations and enhance companies’ intellectual creation cycles. These services will connect R&D departments with IP departments and support the entire process, from initial invention consultations through to the strategic utilization of intellectual property.
As the first offering, NTT DATA plans to provide an AI-powered invention discovery support service that streamlines the process from invention consultation to the preparation of invention disclosure documents.
I asked generative AI to conduct an in-depth analysis of this service, and I invite you to review the results. Please note, however, that the research and analysis generated by AI are based solely on publicly available information. They may not necessarily reflect the actual circumstances and could contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

AI時代の知的財産権検討会(第13回)

20/8/2026

0 Comments

 
2026年8月18日に開催された「AI時代の知的財産権検討会(第13回)」では、生成AI事業者向けの行動指針である「生成AIの適切な利活用等に向けた知的財産の保護及び透明性に関するプリンシプル・コード(仮称)(案)」の修正案を中心に議論が行われました。企業のAI導入やガバナンスのあり方に大きく関わる本コードは、2026年秋をめどに正式版として策定される予定です。
生成AIに、AI時代の知的財産権検討会(第13回)について、内容、反響を調べて、課題、および、企業の知財業務への影響を整理させましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
AI時代の知的財産権検討会(第13回)議事次第
https://www.cas.go.jp/jp/seisakukaigi/titeki2/ai_kentoukai/gijisidai/dai13/index.html
 
 
The 13th Meeting of the Study Group on Intellectual Property Rights in the AI Era
At the 13th meeting of the Study Group on Intellectual Property Rights in the AI Era, held on August 18, 2026, discussions focused primarily on a revised draft of the “Principles Code on Intellectual Property Protection and Transparency for the Appropriate Use of Generative AI” (provisional title), a set of guidelines for generative AI providers. The Code, which will have significant implications for how companies adopt AI and establish AI governance frameworks, is expected to be finalized in the fall of 2026.
I asked generative AI to examine the content of and reactions to the 13th meeting of the Study Group on Intellectual Property Rights in the AI Era, and to identify the key issues and implications for corporate intellectual property operations. Please refer to the resulting analysis.
Please note that the research and analysis conducted by generative AI are based solely on publicly available information. They may not necessarily reflect the actual circumstances and may contain inaccurate information.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

大阪工業大学 杉浦 淳 教授の講演「知的資産と知財教育」

19/8/2026

0 Comments

 
2026年9月4日(金)に、知財・無形資産ガバナンス協会と大阪工業大学の共催による「関西 知財・無形資産ガバナンス・フォーラム KANSAI IP・IA Governance Forum ~関西発 知財・無形資産ドリブン経営への変革~」が、大阪工業大学 梅田キャンパスにて開催されます。申込締め切りが8月20日になっていますので、まだ申し込まれていない方はお急ぎください。
講演④ で、 大阪工業大学 知的財産学部 杉浦 淳 教授が「知的資産と知財教育」というタイトルで講演されますので、生成AIに、杉浦教授のこれまでの業績や論文、講演などを調べて、この講演内容を予測させました。新しい内容が入ってくると思われますのでそもそも予測不能ですが、事前勉強としては最適でしょう。ただ、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。


【知財・無形資産ガバナンス協会×大阪工業大学共催】「関西 知財・無形資産ガバナンス・フォーラム」開催のご案内(9/4開催)
https://chizai-portal.inpit.go.jp/madoguchi/osaka/news/_94.html


Professor Jun Sugiura of Osaka Institute of Technology to Speak on “Intellectual Assets and Intellectual Property Education”
On Friday, September 4, 2026, the KANSAI IP・IA Governance Forum: Transforming Management through Intellectual Property and Intangible Assets—An Initiative from Kansai will be held at the Umeda Campus of Osaka Institute of Technology. The forum is jointly organized by the Intellectual Property and Intangible Asset Governance Association and Osaka Institute of Technology. The registration deadline is August 20, so those who have not yet registered are encouraged to do so as soon as possible.
In Lecture 4, Professor Jun Sugiura of the Faculty of Intellectual Property at Osaka Institute of Technology will deliver a presentation entitled “Intellectual Assets and Intellectual Property Education.”
I therefore asked generative AI to examine Professor Sugiura’s previous achievements, academic papers, lectures, and other activities, and to predict the likely content of his presentation. Since the lecture will presumably include new material, its actual content is, in principle, impossible to predict. Nevertheless, the analysis should provide excellent background preparation before attending the forum.
Please note, however, that the research and analysis conducted by generative AI are based solely on publicly available information. They may not necessarily reflect the actual circumstances and could contain inaccurate information. Please keep this in mind when referring to the analysis.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

クリエイティブな知財部運営に挑むダイキン工業の取組み

18/8/2026

0 Comments

 
2026年9月4日(金)に、知財・無形資産ガバナンス協会と大阪工業大学の共催による「関西 知財・無形資産ガバナンス・フォーラム KANSAI IP・IA Governance Forum ~関西発 知財・無形資産ドリブン経営への変革~」が、大阪工業大学 梅田キャンパスにて開催されます。本フォーラムでは、「知財を経営の力に」をテーマに掲げ、関西から、政府の最新政策、大企業・中小企業の具体的な経営戦略、大学での人財育成などの「知財・無形資産経営」の具体的な実践方法とその将来像を発信します。
講演 ①「知的財産推進計画2026 ~成長戦略を支える知財戦略の推進~」
講演 ②「MPDP理論による中小企業活性化戦略」
講演 ③「経営に活かす知財の在り方を考える ~クリエイティブな知財部運営に挑むダイキン工業の取組み~」
講演 ④「知的資産と知財教育」
パネルディスカッション「知財を経営に活かす人財」のあるべき姿
 
『講演③ 「経営に活かす知財の在り方を考える~クリエイティブな知財部運営に挑むダイキン工業の取組み~」 ダイキン工業株式会社 知的財産部長 安部剛夫 氏』の講演内容を生成AIに予測させましたのでご参照ください。新しい内容が入ってくると思われますのでそもそも予測不能ですが、事前勉強としては最適でしょう。ただ、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
【知財・無形資産ガバナンス協会×大阪工業大学共催】「関西 知財・無形資産ガバナンス・フォーラム」開催のご案内(9/4開催)
https://chizai-portal.inpit.go.jp/madoguchi/osaka/news/_94.html
 
 
Daikin Industries’ Initiatives to Build a Creative Intellectual Property Department
On Friday, September 4, 2026, the KANSAI IP & IA Governance Forum—Transformation toward IP- and Intangible Asset-Driven Management from Kansai will be held at the Umeda Campus of Osaka Institute of Technology. The forum is jointly organized by the Intellectual Property and Intangible Asset Governance Association and Osaka Institute of Technology.
Under the theme “Turning Intellectual Property into a Source of Management Strength,” the forum will present practical approaches to IP and intangible asset management, as well as its future direction, from the Kansai region. Topics will include the Japanese government’s latest policies, concrete management strategies adopted by large corporations and small and medium-sized enterprises, and human resource development at universities.
Presentation 1: “Intellectual Property Strategic Program 2026—Promoting IP Strategies that Support Japan’s Growth Strategy”
Presentation 2: “A Strategy for Revitalizing Small and Medium-Sized Enterprises through MPDP Theory”
Presentation 3: “Rethinking How Intellectual Property Can Contribute to Management—Daikin Industries’ Initiatives to Build a Creative IP Department”
Presentation 4: “Intellectual Assets and Intellectual Property Education”
Panel Discussion: “The Ideal Profile of Talent Capable of Leveraging Intellectual Property in Management”
I asked generative AI to predict the content of Presentation 3, “Rethinking How Intellectual Property Can Contribute to Management—Daikin Industries’ Initiatives to Build a Creative IP Department,” to be delivered by Mr. Takeo Abe, Head of the Intellectual Property Department at Daikin Industries, Ltd.
Because the presentation is expected to contain new information, its actual content is inherently difficult to predict. Nevertheless, the analysis should provide useful background reading and preparation before attending the forum.
Please note, however, that the research and analysis generated by AI are based solely on publicly available information. They may not necessarily reflect the actual circumstances and may contain inaccurate information. Please keep this limitation in mind when referring to the analysis.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

韓国の新興企業が開発した最先端LLM「Motif 3」

17/8/2026

0 Comments

 
韓国の新興企業が開発したMotif 3のArtificial Analysis Intelligence Indexスコアは47で同等のモデルの中では平均を大きく上回る水準で、Motif 3の知能は最先端モデルの一群に入るということで、独自の高度なアーキテクチャを採用することで、この性能を実現しました。この成功の背景には、政府による計算資源の直接提供や、パラメータ効率を極限まで追求する技術戦略といった韓国独自の強力なエコシステムが存在します。
対照的に日本のモデルのグローバルな評価指標での露出は限定的となっています。
生成AIに、限られたリソースで技術的優位を築いた韓国の事例を分析させましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
“Motif 3,” a Frontier LLM Developed by a South Korean Startup
Motif 3, developed by a South Korean startup, achieved a score of 47 on the Artificial Analysis Intelligence Index, significantly outperforming the average among models in its class and placing it among the group of frontier AI models. This level of performance was made possible by the adoption of a distinctive and highly advanced architecture.
Behind this success lies South Korea’s uniquely robust AI ecosystem, including the government’s direct provision of computing resources and a technological strategy focused on pushing parameter efficiency to its limits.
By contrast, Japanese-developed models have had only limited visibility in global AI benchmarks and evaluation rankings.
I asked generative AI to analyze how South Korea has managed to establish a technological advantage despite limited resources. Please refer to the analysis for further details. Please note, however, that the research and analysis produced by generative AI are based solely on publicly available information, may not necessarily reflect the full reality, and may contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

2027年次世代フロンティアAIモデル予測

17/8/2026

0 Comments

 
2026年7月以降に次々とリリースされた最新AIモデルにより、2026年夏時点における米中両国の最新AIモデルの性能差は数パーセントまで縮小しています。
生成AIに、2026年後半から2027年にかけての次世代フロンティアAIモデルにおける技術的展望と、米国・中国間の競争状況を多角的に分析させましたのでご参照ください。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
Forecast for Next-Generation Frontier AI Models in 2027
With a succession of cutting-edge AI models released since July 2026, the performance gap between the latest U.S. and Chinese models has narrowed to just a few percentage points as of summer 2026.
I asked generative AI to conduct a multifaceted analysis of the technological outlook for next-generation frontier AI models from the second half of 2026 through 2027, as well as the evolving competitive landscape between the United States and China. Please take a look.
Please note that the research and analysis generated by AI are based solely on publicly available information. They may not necessarily reflect the actual state of affairs and may contain inaccurate information.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

2026年夏時点における米中両国の最新AIモデル

17/8/2026

0 Comments

 
2026年7月以降に次々とリリースされた最新AIモデルを見ると、かつて言われていた「中国は米国に半年遅れている」という説は、リリース間隔の短縮や高度な推論・プログラミング能力の向上により、現在では実質的に払拭されたと感じます。 2026年夏時点における米中両国の最新AIモデルの技術動向と市場競争を、生成AIに包括的に分析させましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
The Latest U.S. and Chinese AI Models as of Summer 2026
Looking at the latest AI models released in rapid succession since July 2026, the once-common view that “China is six months behind the United States” now appears to have been effectively dispelled, given the shortening of release cycles and the substantial improvements in advanced reasoning and programming capabilities.
I asked generative AI to conduct a comprehensive analysis of the technological trends and market competition surrounding the latest AI models from the United States and China as of summer 2026. Please refer to the analysis for further details.
Please note, however, that the research and analysis conducted by generative AI are based solely on publicly available information. They may not necessarily reflect the actual situation and may contain inaccurate information.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

生成AI活用特許データ分析による事業機会探索

17/8/2026

0 Comments

 
川上成年氏の提言「生成 AI を活用した特許データ分析による事業機会探索手法の開発(月刊「パテント」2026年 6月)」は、大規模言語モデル(LLM)を特許データの解析に導入することで、新たな収益源となる事業機会を効率的に特定するフレームワークを提案しています。従来は分断されていた数値的な定量分析と文脈を読み解く定性分析を、生成AIという共通基盤上でシームレスに統合し、最終的な事業コンセプトの創出まで一気通貫で行っています。mRNA医薬を対象としたケーススタディを通じて、この手法が従来の膨大なコストや時間を大幅に削減し、客観的かつ迅速な意思決定を可能にすることを実証しています。
この提言について、生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
生成AIを活用した特許データ分析による事業機会探索手法の開発
月刊「パテント」Vol. 79 ,No. 6, P.55-P.65 (2026)
https://jpaa-patent.info/patent/viewPdf/4837
 
 
Exploring Business Opportunities through Generative AI–Powered Patent Data Analysis
Naritoshi Kawakami’s article, “Development of a Method for Exploring Business Opportunities through Patent Data Analysis Using Generative AI,” published in the June 2026 issue of the monthly journal Patent, proposes a framework for efficiently identifying business opportunities that could become new sources of revenue by applying large language models (LLMs) to patent data analysis.
The proposed method seamlessly integrates quantitative analysis based on numerical data with qualitative analysis that interprets contextual meaning—two processes that have traditionally been conducted separately—on a common generative AI platform. It then carries the analysis through, in an end-to-end manner, to the creation of concrete business concepts. Through a case study involving mRNA therapeutics, the article demonstrates that this approach can substantially reduce the enormous costs and time traditionally required while enabling faster and more objective decision-making.
I asked generative AI to conduct an in-depth analysis of this article and invite you to review the results. Please note, however, that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect actual circumstances, and may contain inaccurate information.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

ジェフ・ディーン氏らが新会社「Discovery Loop」を設立

17/8/2026

0 Comments

 
現代コンピュータサイエンスの重鎮であるジェフ・ディーン氏が、27年間在籍したGoogleを退職し、サンジェイ・ゲマワット氏、クォック・リー氏、オリオール・ビニャルス氏とともに、新会社Discovery Loopを設立したことが報じられています。
この新会社は、AIによる科学研究プロセスの完全自動化を掲げており、仮説立案から実験、評価までを自律的に高速化することを目指しています。共同創業者には分散システムや機械学習の世界的権威が名を連ね、Alphabetも戦略的投資家として支援する体制が構築されました。
この動きは、単純な「Googleからの頭脳流出」とは言い切れません。Alphabetは創業投資家として参加し、Google Cloudや計算資源のパートナーとなるほか、機械学習システムに関する研究協力も継続します。Googleが希少な人材を失う一方、独立組織による長期的・高リスクな研究の成果を取り込む選択肢を残した「協調型独立」あるいは「戦略的ヘッジ」と見ることもできます。
このDiscovery Loopについて、生成AIに多面的に深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
Jeff Dean and Fellow Researchers Establish New Company, Discovery Loop
Jeff Dean, one of the most influential figures in modern computer science, has reportedly left Google after 27 years to establish a new company, Discovery Loop, together with Sanjay Ghemawat, Quoc Le, and Oriol Vinyals.
The new company aims to achieve end-to-end automation of the scientific research process through AI, accelerating the entire cycle—from hypothesis generation to experimentation and evaluation—through autonomous systems. Its co-founders include world-renowned authorities in distributed systems and machine learning, while Alphabet has joined as a strategic investor.
This development cannot simply be described as a “brain drain” from Google. Alphabet is participating as a founding investor, Google Cloud and other computing resources are expected to support the company’s operations, and research collaboration in machine-learning systems will reportedly continue. Although Google is losing exceptionally rare talent, the arrangement also leaves it with an opportunity to benefit from the results of long-term, high-risk research conducted by an independent organization. In this sense, the move may be better understood as a form of “collaborative independence” or a “strategic hedge.”
I asked generative AI to conduct an in-depth, multifaceted analysis of Discovery Loop. Please refer to the resulting report. Please note, however, that the research and analysis generated by AI are based solely on publicly available information and may not necessarily reflect the actual circumstances. They may also contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

高性能AIの悪用に備えた知財保護ルール

16/8/2026

0 Comments

 
政府は高性能AIの普及を踏まえ省庁横断で包括的に対策を講じる、という記事が出ていました。この件を、生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
AI悪用に備え知財保護ルール 政府、省庁横断で検討
2026年8月14日
https://www.nikkei.com/article/DGKKZO98142170T10C26A8EA1000/
 
 
Rules for Protecting Intellectual Property Against the Misuse of Advanced AI
An article reported that, in response to the growing adoption of advanced AI, the Japanese government is preparing to implement comprehensive, cross-ministerial measures. I asked a generative AI system to examine this issue in greater depth, and the resulting analysis is provided for your reference.
Please note that the research and analysis conducted by the generative AI are based solely on publicly available information, may not necessarily reflect the actual circumstances, and may contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

Anthropic第2回リスクレポートが示す、これからの知財実務

16/8/2026

0 Comments

 
Anthropicは2026年8月14日、186ページに及ぶ第2回「Risk Report」を公開しました。本報告書は、AIモデルの能力だけを評価するものではなく、監視、アクセス制御、セキュリティ、導入時の安全措置まで含めて、同社のAI開発・利用活動全体がもたらす破局的リスクを検討したものです。主な対象は、高リスク環境におけるミスアラインメント、AIによる研究開発の自動化、化学・生物兵器の開発・製造などです。
各領域の総合評価はおおむね「Low」とされています。しかし、その一方で、AIの危険な行動を測定する評価手法の限界、研究開発を加速させる初期的な兆候、アクセス管理上の不備、監視から漏れる利用領域など、不確実性や安全対策の弱点も率直に記載されています。したがって、この「Low」を、「企業における日常的なAI利用も安全である」という意味に読み替えることはできません。
知財実務では、未公開発明、FTO調査資料、訴訟・M&A関連資料、営業秘密などをAIに入力する場面や、AIエージェントに文書の作成・変更・送信権限を与える場面が増えています。こうした利用は、破局的リスクに至らなくても、秘密管理性の喪失、発明者認定や権利帰属をめぐる争い、AI生成物の出典・根拠の不明確化、調査・分析結果の証拠力低下といった、知財固有の重大な問題を引き起こす可能性があります。AIを使う場合に、「何を、どの証拠で、どこまで任せるか」を設計することの重要性がより高まっているといえます。
生成AIに、Anthropicのリスクレポートが今後の知財実務に及ぼす影響を、機密情報管理、発明者と人間の創作的寄与の記録、AI生成調査の証拠化、AIエージェントの権限設計、ベンダー契約や監査条項などへの影響を多面的に検討させましたので、ご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
Anthropic「Risk Report: August 2026」
https://www-cdn.anthropic.com/f61d49fa5596956a5dec75fea0e973bf6a6a8378/Redacted%20Risk%20Report%20August%202026%20.pdf
 
 
What Anthropic’s Second Risk Report Means for the Future of Intellectual Property Practice
On August 14, 2026, Anthropic published its second Risk Report, a 186-page assessment. Rather than evaluating only the capabilities of its AI models, the report examines the potential catastrophic risks arising from the full range of the company’s AI development and deployment activities, including monitoring, access controls, security, and safeguards implemented at the time of deployment. Its principal areas of focus include misalignment in high-risk environments, AI-driven automation of research and development, and the development and production of chemical and biological weapons.
The overall risk rating for each area is generally assessed as “Low.” At the same time, however, the report candidly acknowledges substantial uncertainties and weaknesses in existing safeguards, including the limitations of evaluation methods for measuring dangerous AI behavior, early indications that AI may accelerate research and development, deficiencies in access controls, and areas of use that may escape monitoring. Accordingly, the “Low” rating should not be interpreted as meaning that routine corporate use of AI is also inherently safe.
In intellectual property practice, AI is increasingly being used to process non-public inventions, freedom-to-operate search materials, litigation- and M&A-related documents, trade secrets, and other sensitive information. AI agents are also increasingly being granted authority to create, modify, and transmit documents. Even when such uses do not give rise to catastrophic risks, they may still create serious IP-specific problems, including the loss of trade-secret protection due to inadequate confidentiality controls, disputes over inventorship and ownership of rights, uncertainty regarding the provenance and evidentiary basis of AI-generated content, and a decline in the evidentiary weight of AI-generated search and analysis results. This makes it increasingly important to design, in advance, what tasks should be delegated to AI, what evidence should support its outputs, and how far its authority should extend.
I asked generative AI to conduct a multifaceted examination of how Anthropic’s Risk Report may affect future intellectual property practice, including its implications for confidential information management, the documentation of inventorship and human creative contributions, the preservation of AI-generated research as evidence, the design of permissions and authority for AI agents, and vendor agreements and audit provisions. Please refer to the resulting analysis. Please note, however, that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect actual circumstances, and may contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

脆弱性の発見性能「ミュトス級」の「GLM-5.3」

16/8/2026

0 Comments

 
中国のAI企業Z.ai(旧 智譜AI)は、2026年8月14日に最新の大規模言語モデル「GLM-5.3」を発表・リリースしました。前バージョンのGLM-5.2と同じ7,430億パラメータのベースモデルをそのまま使用しながら、「事後学習」の大幅な拡張だけで性能向上を達成し、脆弱性の特定・検証能力を測定するベンチマーク「CyberGym」では、84.5%のスコアで、Fable 5やGPT-5.6 Sol、Kimi K3などの競合モデルを上回りました。
この「GLM-5.3」について、生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
Fable 5とも戦える「GLM-5.3」登場。事後学習の拡張のみで大幅性能アップ
2026年8月14日
https://pc.watch.impress.co.jp/docs/news/2132889.html#04_l.jpg
 
中国Z.AI、脆弱性の発見性能「ミュトス級」 新モデル提供開始
2026年8月14日
https://www.nikkei.com/article/DGXZQOGM144JU0U6A810C2000000/
 
 
GLM-5.3 Achieves “Mythos-Class” Vulnerability Discovery Performance
On August 14, 2026, Chinese AI company Z.ai, formerly known as Zhipu AI, announced and released its latest large language model, GLM-5.3. While retaining the same 743-billion-parameter base model used in the previous version, GLM-5.2, the company achieved substantial performance gains solely by significantly expanding its post-training process.
On CyberGym, a benchmark that measures models’ ability to identify and validate software vulnerabilities, GLM-5.3 scored 84.5%, outperforming competing models such as Fable 5, GPT-5.6 Sol, and Kimi K3.
I asked generative AI to conduct an in-depth analysis of GLM-5.3. Please refer to the results for further details. Please note, however, that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect the actual situation, and may contain inaccurate information.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

デュアルユース技術の最新動向

15/8/2026

0 Comments

 
かつての軍から民への技術転用とは対照的に、現代では民生技術が軍事に取り込まれる「スピンオン」が主流となっており、AIやドローン、宇宙開発がその中核を担っています。米国が制度改革や巨額の投資で産業基盤を強化する一方、日本も防衛予算の拡大や新研究機関の設立、さらに学術界の姿勢変化を背景に、大きな転換期を迎えています。
生成AIに、軍事と民間の両面で活用されるデュアルユース技術の最新動向を、日本と米国の比較を含め詳細に分析させましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
 
Latest Trends in Dual-Use Technologies
In contrast to the traditional transfer of military technologies to civilian applications, today the dominant trend is “spin-on,” in which civilian technologies are incorporated into military systems. Artificial intelligence, drones, and space technologies are at the core of this shift. While the United States is strengthening its defense industrial base through institutional reforms and massive investment, Japan is also entering a major period of transition, driven by an expanding defense budget, the establishment of new research organizations, and changing attitudes within the academic community.
I asked generative AI to conduct a detailed analysis of the latest developments in dual-use technologies employed for both military and civilian purposes, including a comparison between Japan and the United States. Please refer to the analysis for further details. Please note, however, that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect the actual circumstances, and may contain inaccuracies.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments

Gemini 3.7 Flash

15/8/2026

0 Comments

 
Googleは2026年8月13日(米国時間)、コーディングやエージェント機能を強化した主力モデル「Gemini 3.7 Flash」を公開しました。前バージョンのGemini 3.6 Flashから3週間で新モデルが登場したこととなります。
Gemini 3.7 Flashは、ソフトウェア エンジニアリング、ナレッジ ワーク、Web開発など、ワークフロー全体で大幅に進化したとされ、100万トークンあたりの費用はGemini 3.6 Flashの「半額」の初回価格に抑えています。
このGemini 3.7 Flashについて生成AIに深掘りさせましたのでご参照ください。なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
Gemini 3.7 Flash公開 主力モデルを3週間で更新、性能も大幅向上
2026年8月14日
https://www.watch.impress.co.jp/docs/news/2132694.html
 
 
Gemini 3.7 Flash
On August 13, 2026 (U.S. time), Google released Gemini 3.7 Flash, its flagship model with enhanced coding and agentic capabilities. The new model arrived just three weeks after the release of its predecessor, Gemini 3.6 Flash.
Gemini 3.7 Flash is said to deliver major improvements across entire workflows, including software engineering, knowledge work, and web development. Its introductory price per one million tokens has also been set at half that of Gemini 3.6 Flash.
I asked generative AI to conduct an in-depth analysis of Gemini 3.7 Flash, and I hope you will find the results informative. Please note, however, that the research and analysis produced by generative AI are based solely on publicly available information and may not necessarily reflect the actual situation. They may also contain inaccuracies, so please bear this in mind when reviewing the material.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments
<<Previous
Forward>>

    著者

    萬秀憲

    アーカイブ

    April 2026
    March 2026
    February 2026
    January 2026
    December 2025
    November 2025
    October 2025
    September 2025
    August 2025
    July 2025
    June 2025
    May 2025
    April 2025
    March 2025
    February 2025
    January 2025
    December 2024
    November 2024
    October 2024
    September 2024
    August 2024
    July 2024
    June 2024
    May 2024
    April 2024
    March 2024
    February 2024
    January 2024
    December 2023
    November 2023
    October 2023
    September 2023
    August 2023
    July 2023
    June 2023
    May 2023
    April 2023
    March 2023
    February 2023
    January 2023
    December 2022
    November 2022
    October 2022
    September 2022
    August 2022
    July 2022
    June 2022
    May 2022
    April 2022
    March 2022
    February 2022
    January 2022
    December 2021
    November 2021
    October 2021
    September 2021
    August 2021
    July 2021
    June 2021
    May 2021
    April 2021
    March 2021
    February 2021
    January 2021
    December 2020
    November 2020
    October 2020
    September 2020
    August 2020
    July 2020
    June 2020

    カテゴリー

    All

    RSS Feed

Copyright © よろず知財戦略コンサルティング All Rights Reserved.
サイトはWeeblyにより提供され、お名前.comにより管理されています
  • Home
  • Services
  • About
  • Contact
  • Blog
  • 知財活動のROICへの貢献
  • 生成AIを活用した知財戦略の策定方法
  • 生成AIとの「壁打ち」で、新たな発明を創出する方法