• 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

不正競争防止法に基づく訴訟提起

19/9/2026

0 Comments

 
パテント・インテグレーションが、2026年9月14日、「京都大学発 ・松尾研発スタートアップ」として、製造業向けの生成AI開発等を手がける、株式会社エムニに対し、同社が提供するサービス「AI特許ロケット」の広告表示に関し、不正競争防止法に基づく差止等請求訴訟を東京地方裁判所に提起したということです。「外注時と比較して99.9%のコスト削減」「業界破壊の低価格」という表示が問題となっています。
生成AIに、本件について深堀させましたので、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。
 
株式会社エムニに対する不正競争防止法に基づく訴訟提起に関するお知らせ
パテント・インテグレーション株式会社
2026年9月15日
https://prtimes.jp/main/html/rd/p/000000030.000086119.html
 
 
Lawsuit Filed Under the Unfair Competition Prevention Act
On September 14, 2026, Patent Integration reportedly filed a lawsuit in the Tokyo District Court against Emuni Inc., seeking an injunction and other relief under Japan’s Unfair Competition Prevention Act over advertising claims for Emuni’s “AI Patent Rocket” service. Emuni is described as a “startup originating from Kyoto University and the Matsuo Lab” and develops generative AI solutions and related technologies for the manufacturing sector. The advertising claims apparently at issue include “99.9% cost reduction compared with outsourcing” and “industry-disrupting low prices.”
I asked generative AI to conduct an in-depth investigation into this case. The findings are provided for your reference.
Please note that these AI-generated research and analysis results are based solely on publicly available information, do not necessarily reflect the actual situation, and may contain inaccuracies. Please bear these limitations in mind when reviewing the findings.

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

すでにAGIは達成されたのか?

19/9/2026

0 Comments

 
NVIDIAのジェンスン・フアンCEOはLex Fridmanのインタビューで「AGIを達成したと考えている」と述べ、OpenAIのグレッグ・ブロックマン社長も2026年9月3日の会見で「AGIの時代へようこそ」と発言しました。しかし、一般利用者が触れているのは製品化されたAIであり、研究者が最前線で観察している対象とは性質が異なります。日常的な利用では依然として多くの課題や限界が見受けられるため、「すでにAGIが達成された」という認識にはまだ実感が伴っていないのが実情です。
生成AIにこの辺りの事情を深掘り調査させましたので、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。
 
Has AGI Already Been Achieved?
NVIDIA CEO Jensen Huang stated in an interview with Lex Fridman, “I believe we have achieved AGI,” while OpenAI President Greg Brockman said, “Welcome to the era of AGI,” at a press conference on September 3, 2026. However, ordinary users interact with AI systems released as products, which differ in nature from what researchers are observing at the forefront of AI development. With many challenges and limitations still apparent in everyday use, the notion that “AGI has already been achieved” has yet to feel like a reality for ordinary users.
I asked generative AI to conduct an in-depth investigation into the circumstances behind this gap in perception. Please refer to the findings.
Please note, however, that these AI-generated research and analysis results are based solely on publicly available information, do not necessarily reflect the actual state of affairs, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

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

すでにAGIは達成されたのか?

19/9/2026

0 Comments

 
NVIDIAのジェンスン・フアンCEOはLex Fridmanのインタビューで「AGIを達成したと考えている」と述べ、OpenAIのグレッグ・ブロックマン社長も2026年9月3日の会見で「AGIの時代へようこそ」と発言しました。しかし、一般利用者が触れているのは製品化されたAIであり、研究者が最前線で観察している対象とは性質が異なります。日常的な利用では依然として多くの課題や限界が見受けられるため、「すでにAGIが達成された」という認識にはまだ実感が伴っていないのが実情です。
生成AIにこの辺りの事情を深掘り調査させましたので、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。
 
 
Has AGI Already Been Achieved?
NVIDIA CEO Jensen Huang stated in an interview with Lex Fridman, “I believe we have achieved AGI,” while OpenAI President Greg Brockman said, “Welcome to the era of AGI,” at a press conference on September 3, 2026. However, ordinary users interact with AI systems released as products, which differ in nature from what researchers are observing at the forefront of AI development. With many challenges and limitations still apparent in everyday use, the notion that “AGI has already been achieved” has yet to feel like a reality for ordinary users.
I asked generative AI to conduct an in-depth investigation into the circumstances behind this gap in perception. Please refer to the findings.
Please note, however, that these AI-generated research and analysis results are based solely on publicly available information, do not necessarily reflect the actual state of affairs, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

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年知財・情報フェア&コンファレンス

18/9/2026

0 Comments

 
2026年9月16日から行われている「2026年知財・情報フェア&コンファレンス」では、生成AIや大規模言語モデルを活用したサービスが主流となり、単なる書類作成の効率化を超え、経営判断や事業価値の創出に直結する仕組みが数多く提案されています。特に、社内データとの連携や特許・商標・ブランド保護の統合管理など、組織全体で情報を再利用する実務的な設計が注目されています。また、AI導入による工数削減だけでなく、効率化で生まれた時間を人材育成や戦略立案に充てるという、知財部門の役割の変化、最新技術をいかに企業の競争力強化や意思決定に結びつけるかという、具体的かつ高度な実装段階への移行が提案されていました。
「2026年知財・情報フェア&コンファレンス」における展示や講演内容を、生成AIに深掘り調査させましたので、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。
 
 
2026 Patent Information Fair & Conference
At the 2026 Patent Information Fair & Conference, which opened on September 16, 2026, services leveraging generative AI and large language models have taken center stage. Numerous solutions are being proposed that go beyond simply streamlining document preparation to directly support management decision-making and the creation of business value. Particular attention is being paid to practical approaches that enable information to be reused across the organization, including integration with internal corporate data and the unified management of patents, trademarks, and brand protection. The exhibits and presentations also proposed a shift in the role of IP departments: moving beyond simply reducing workloads through AI adoption to devoting the time saved to talent development and strategic planning. They also advocated a transition toward a more concrete and sophisticated phase of implementation, focusing on how the latest technologies can be translated into stronger corporate competitiveness and better decision-making.
I asked generative AI to conduct an in-depth investigation into the exhibits and presentations at the 2026 Patent Information Fair & Conference, and am sharing the findings for your reference. Please note that these AI-generated research and analysis results are based solely on publicly available information, do not necessarily reflect the actual situation, and may contain errors. Please review them with these limitations in mind.

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 for Science の進化と知財業務への影響

16/9/2026

0 Comments

 
「AI for Scienceとは何か――AIが「科学者を助ける」段階から「科学そのものを実行する」段階へ」という石川 有一さんの投稿に触発され、AI for Science の進化と知財業務への影響について生成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。
 
AI for Scienceとは何か――AIが「科学者を助ける」段階から「科学そのものを実行する」段階へ
https://www.facebook.com/groups/548266103946448/permalink/1401824511923932/?rdid=4JI3PU177JhsaXsb#
 
 
The Evolution of AI for Science and Its Impact on IP Work
Inspired by Yuichi Ishikawa’s post, “What Is AI for Science?—From AI That ‘Helps Scientists’ to AI That ‘Conducts Science Itself’,” I asked generative AI to conduct an in-depth exploration of the evolution of AI for Science and its impact on intellectual property (IP) work. Please take a look at the findings.
Please note that the research and analysis produced by generative AI are based solely on publicly available information, may not necessarily reflect actual circumstances, and may contain errors. Please bear these limitations in mind when reviewing the results.

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

PerplexityのローカルAIエージェント「Portable Computer」

16/9/2026

0 Comments

 
NVIDIAは9月14日(現地時間)、PerplexityのローカルAIエージェント「Portable Computer」が「NVIDIA RTX」GPUを搭載したWindows PCで利用可能になったと発表しました。今後、企業でのローカルLLMの構築が加速し、フロンティアLLMとの併用が進むと思われます。
PerplexityのローカルAIエージェント「Portable Computer」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。
 
 
Perplexity Portable Computer Is Now Available on Windows, Powered by NVIDIA RTX
https://blogs.nvidia.com/blog/local-ai-perplexity-windows-pcs/
 
Windows と Linux 対応の Portable Computerローカルファーストのエージェントスタック
https://www.perplexity.ai/hub/products/portable-computer
 
「Perplexity」のローカルAIエージェント「Portable Computer」がWindowsでも利用可能に、VRAM 24GB以上のNVIDIA RTX GPUで
2026年9月15日
https://forest.watch.impress.co.jp/docs/news/2140871.html
 
 
Perplexity’s Local AI Agent “Portable Computer”
On September 14 (local time), NVIDIA announced that Perplexity’s local AI agent, “Portable Computer,” is now available on Windows PCs equipped with NVIDIA RTX GPUs. Looking ahead, enterprises are likely to accelerate the deployment of local large language models (LLMs) and increasingly use them alongside frontier LLMs.
I asked generative AI to conduct an in-depth investigation into Perplexity’s local AI agent, “Portable Computer.” Please see the findings for your reference.
Please note that these AI-generated research and analysis findings are based solely on publicly available information, do not necessarily reflect actual circumstances, and may contain inaccuracies. Please keep these limitations in mind when reviewing them.

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

汎用人工知能・AGIはもう来ている

16/9/2026

0 Comments

 
Forbesの『「汎用人工知能・AGIはもう来ている」量子物理学者が告げた理由』という記事では、Haiqu共同創業者兼CTOで量子物理学者であるMaksymenko氏が、AIエージェント・LLM・現行量子コンピューターの組み合わせによって自身の博士研究が週末で再現・修正された体験に衝撃を受け、社内会議で「AGIはもう来ていると思う」と発言したエピソードを紹介しています。
この記事について生成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。
 
「汎用人工知能・AGIはもう来ている」量子物理学者が告げた理由
9/14(月)
https://news.yahoo.co.jp/articles/cb5f3053a5709930e0af3c5d66dd258c2210217f
 
 
Artificial General Intelligence (AGI) Is Already Here
A Forbes article titled “‘Artificial General Intelligence (AGI) Is Already Here’: Why a Quantum Physicist Says So” recounts how Maksymenko, a quantum physicist and co-founder and CTO of Haiqu, was stunned when a combination of AI agents, large language models (LLMs), and current quantum computers reproduced and corrected his doctoral research over a single weekend. At an internal company meeting, he remarked, “I think AGI is already here.”
I asked generative AI to explore this article in greater depth. Please see the resulting research and analysis.
Please note that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect the actual state of affairs, and may contain inaccuracies. Please bear these limitations in mind when reviewing the results.

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

米国で空前の起業ブーム「自分で会社」

15/9/2026

0 Comments

 
米国では、人工知能(AI)の普及や雇用への不安を背景に、若者が就職よりも「自分で会社を持つ」道を選ぶ空前の起業ブームが起きています。若者の就職事情について、生成AIに調べさせ、米国と日本との違いを考察させましたので、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。
 
米国で空前の起業ブーム AIと若者、ブルーカラーより「自分で会社」
2026年9月10日
https://www.nikkei.com/article/DGXZQOGN27BAF0X20C26A8000000/
 
米国で空前の起業ブーム!AIと若者が目指す「自分で会社を持つ」という生き方
2026/09/10
https://know-how-tree.com/archives/31464
 
 
An Unprecedented Entrepreneurship Boom in the United States
Against a backdrop of growing adoption of artificial intelligence (AI) and concerns about employment prospects, the United States is experiencing an unprecedented entrepreneurship boom, with young people choosing to “own their own businesses” rather than seek employment. I asked generative AI to investigate the job market facing young people and analyze the differences between the United States and Japan. The findings are presented here for your reference.
Please note that the research and analysis generated by AI are based solely on publicly available information, may not necessarily reflect actual conditions, and may contain errors. Please bear these limitations in mind when reviewing the findings.

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

サポート要件違反に対する反論についての検討

15/9/2026

0 Comments

 
生成AIの性能が急速に向上しています。専門家の論説を批評させることで、そのレベルを把握すべく、月刊パテント 2026年7月号に掲載されている、日本弁理士会令和6年度特許委員会第2部会第2グループによる活動報告「サポート要件違反に対する反論についての検」について、生成AIに批評させましたので、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。
 
日本弁理士会・令和6年度特許委員会第2部会第2グループ(田中研二ほか)「サポート要件違反に対する反論についての検討」『パテント』79巻7号、2026年、38–45頁
https://jpaa-patent.info/patent/viewPdf/4851
 
 
A Study of Rebuttals to Alleged Noncompliance with the Support Requirement
The capabilities of generative AI are improving rapidly. To assess its ability to critically evaluate articles written by experts, I asked generative AI to critique the activity report “A Study of Rebuttals to Alleged Noncompliance with the Support Requirement,” prepared by Group 2 of the Second Subcommittee of the Japan Patent Attorneys Association’s FY2024 Patent Committee and published in the July 2026 issue of the monthly journal Patent. Please see the resulting critique.
Please note 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 errors. Please bear these limitations in mind when reading the critique.

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

分割出願の活用事例についての検討

14/9/2026

0 Comments

 
生成AIの性能が急速に向上しています。専門家の論説を批評させることで、そのレベルを把握すべく、月刊パテント 2026年7月号に掲載されている、日本弁理士会令和6年度特許委員会第2部会第1グループによる活動報告「分割出願の活用事例についての検討」について、生成AIに批評させましたので、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。
 
分割出願の活用事例についての検討
https://jpaa-patent.info/patent/viewPdf/4850
 
 
A Study of Use Cases for Divisional Applications
The capabilities of generative AI are improving rapidly. To assess its ability to critically evaluate expert commentary, I asked generative AI to critique “A Study of Use Cases for Divisional Applications,” an activity report prepared by Group 1 of Subcommittee 2 of the Japan Patent Attorneys Association’s FY2024 Patent Committee and published in the July 2026 issue of the monthly journal Patent. Please see the resulting critique.
Please note that the research and analysis produced by generative AI are based solely on publicly available information, do not necessarily reflect actual circumstances, and may contain inaccuracies. Please bear these limitations in mind when reading the review.

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による自律的な発明・研究能力の評価が本格化

14/9/2026

0 Comments

 
ARC Prizeは、2026年9月12日、自律的なオープンエンドのイノベーションを対象にする予定のベンチマークである「ARC-AGI-4」を将来版として予告しました。2027年初頭の公開が予定されていますが、具体的な評価設計はまだ公表されていません。ARC-AGI-4が「終わりなき発明(open-ended invention)」「再帰的自己改善」を測定対象に据える方向性は、AIによる自律的な発明・研究能力の評価が本格化することを意味すると考えられます。
「ARC-AGI-4」を巡る動きについて、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。
 
ARC-AGI-4 will be a benchmark for autonomous open-ended innovation
https://x.com/arcprize/status/2098849962754978152
 
 
Efforts to Evaluate AI’s Autonomous Invention and Research Capabilities Gain Momentum
On September 12, 2026, ARC Prize announced plans for ARC-AGI-4, a future benchmark intended to assess autonomous, open-ended innovation. Its release is scheduled for early 2027, but the detailed evaluation design has not yet been made public. The proposed direction of ARC-AGI-4—assessing “open-ended invention” and “recursive self-improvement”—suggests that efforts to evaluate AI’s capacity for autonomous invention and research are beginning in earnest.
I asked generative AI to conduct an in-depth investigation into developments surrounding ARC-AGI-4. Please refer to the resulting analysis. When reviewing it, please bear in mind 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 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

PatRe:特許審査全工程をモデル化した新評価指標

13/9/2026

0 Comments

 
2026年5月5日に公表された「PatRe: A Full-Stage Office Action and Rebuttal Generation Benchmark for Patent Examination (PatRe: 特許審査におけるフルステージのオフィスアクションおよび反論書作成のベンチマーク)」は、複雑な特許審査の全工程をモデル化した新しい評価指標であるPatReを紹介する研究論文です。従来のAI評価が単純な分類に留まっていたのに対し、本論文は審査官による拒絶理由通知(Office Action)と出願人による反論(Rebuttal)のやり取りを、動的な対話プロセスとして再現しています。
GPT-5-mini and GPT-4o-mini, Gemini-2.5-Flash, DeepSeek-V3.2等の大規模言語モデル(LLM)を用いた実験の結果、モデルは既存の議論に応答する能力には長けているものの、自ら不備を発見する能力や法的・技術的な妥当性の判断には依然として課題があることが浮き彫りになりました。このベンチマークは、専門性の高い法的推論におけるAIの限界を明示し、知財分野におけるより高度なモデル開発を促進することを目的としていますので、GPT-6 Astra等の最新のモデルで同じ実験を行えば、成績は改善する可能性が高く、特に「審査官役」と「反論書作成」の差が縮まるのではないかと考えられ、そうした検討結果の報告が期待されます。
本件について、生成AIに深掘り調査させましたので、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。
 
PatRe: A Full-Stage Office Action and Rebuttal Generation Benchmark for Patent Examination (PatRe: 特許審査におけるフルステージのオフィスアクションおよび反論書作成のベンチマーク)
https://arxiv.org/abs/2605.03571
 
 
PatRe: A New Benchmark Modeling the Entire Patent Examination Process
Published on May 5, 2026, the research paper “PatRe: A Full-Stage Office Action and Rebuttal Generation Benchmark for Patent Examination” introduces PatRe, a new benchmark that models the entire complex patent examination process. Whereas previous AI evaluations were limited to simple classification tasks, this paper recreates the exchanges between office actions issued by patent examiners and rebuttals submitted by applicants as a dynamic dialogue.
Experiments using large language models (LLMs), including GPT-5-mini, GPT-4o-mini, Gemini-2.5-Flash, and DeepSeek-V3.2, revealed that although these models are adept at responding to existing arguments, they still face challenges in independently identifying deficiencies and assessing legal and technical validity. The benchmark aims to highlight the limitations of AI in highly specialized legal reasoning and encourage the development of more advanced models for the intellectual property field.
Repeating the same experiments with the latest models, such as GPT-6 Astra, would likely yield improved results. In particular, the performance gap between acting as a patent examiner and drafting applicants’ rebuttals may narrow. Reports examining these possibilities are therefore eagerly awaited.
I asked generative AI to conduct an in-depth investigation of this topic. Please refer to the findings, bearing in mind that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect actual circumstances, 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.
0 Comments

GPT-6 Astraはなぜ評価されるのか

13/9/2026

0 Comments

 
GPT-6 Astraはなぜ評価されるのか?これまでのAIが「答えるAI」だったのが、GPT-6 Astraは「仕事を終わらせるAI」へ進化したからということで、Aino使い方、AIを評価する基準が、「うまく答えるか」から「仕事を最後までやり切れるか」へとシフトしているということのようです。
従来の「1問1答」と違い、ファイル探索・テスト・ログ解析を自律的に繰り返すエージェント処理は、莫大な計算資源を長時間消費します。月額200ドルのChatGPT Proが新規受付を一時停止したのは、新規使用者の急増というより、「チャット」から「長時間の連続実行」へ使い方の変化による可能性が大きいようです。
サービス提供から1週間経過した段階でのGPT-6 Astraの評価を、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。
 
Why Is GPT-6 Astra Earning Praise?
Why is GPT-6 Astra receiving such positive reviews? The reason appears to be that, whereas earlier AI systems were focused on “answering questions,” GPT-6 Astra has evolved into an AI that “gets the job done.” Both the way AI is used and the criteria for evaluating it seem to be shifting from “How well does it answer?” to “Can it see the job through to completion?”
Unlike traditional “one question, one answer” interactions, agentic workflows that autonomously repeat file searches, testing, and log analysis consume enormous computing resources over extended periods. The temporary suspension of new sign-ups for the $200-per-month ChatGPT Pro plan appears more likely to be attributable to a shift in usage—from “chat” to “long-running, continuous execution”—than to a surge in new users alone.
For your reference, I asked generative AI to conduct an in-depth review of the reception and evaluation of GPT-6 Astra one week after its launch. Please note that this AI-generated research and analysis is based solely on publicly available information, may not necessarily reflect the actual situation, and may contain errors. Please review the findings with these limitations in mind.

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

GPT-6 Astraユーザー急増でChatGPT Pro新規受付一時停止

12/9/2026

0 Comments

 
2026年9月11日、OpenAIは新モデルGPT-6 Astraのユーザーの急増により、月3万円のChatGPT Pro新規受付を一時停止しました。新モデルGPT-6 Astraの需要急増に伴い、システムへの負担がかつてない規模に達していることが要因ということです。
について、生成AIに深掘り調査させましたので、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。
 
Astra人気でOpenAIが悲鳴!月3万円のChatGPT Pro新規受付を一時停止
9/11
https://news.yahoo.co.jp/articles/d93c105d6c09d2857176d0a6ce6ce016f2a21292
 
 
OpenAI Temporarily Suspends New ChatGPT Pro Sign-Ups Amid a Surge in GPT-6 Astra Users
On September 11, 2026, OpenAI temporarily suspended new sign-ups for ChatGPT Pro, its ¥30,000-per-month subscription plan, following a surge in users of its new GPT-6 Astra model. The suspension was reportedly prompted by unprecedented strain on its systems as demand for the new model soared.
I asked generative AI to conduct an in-depth investigation into this development and am sharing the findings for your reference. Please note, however, that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect the actual situation, 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.
0 Comments

「仮説を作るAI」出現で変わる科学者の役割

12/9/2026

0 Comments

 
2026年8月22日付の朝日新聞朝刊は、「創薬・数学、AIが新発見 論理的能力磨き、急速に進化」という見出しで、最先端AIが人間には届かなかった科学や数学の分野で新しい発見を成し遂げ始めている現状を一面トップで報じました。AIによる科学的発見の自動化が進み膨大な仮説が供給される時代においては、科学者の役割は「仮説の創出」から「検証プロセスの設計と統治」へと移行するだろうといわれています。
本件について、生成AIに深掘り調査させましたので、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。
 
「仮説を作るAI」出現で科学者たちは不要に!?AIとの“知的分業”をどう進めるのか
9/12(土) 野口悠紀雄 一橋大学名誉教授
https://news.yahoo.co.jp/articles/6b7b48f6a8e6e050ed6212cd659fb9c2b08b2341
 
(AIの時代)創薬・数学、AIが新発見 論理的能力磨き、急速に進化
2026年8月22日
https://digital.asahi.com/articles/DA3S16531060.html
 
 
How the Emergence of “Hypothesis-Generating AI” Is Changing the Role of Scientists
The August 22, 2026 morning edition of The Asahi Shimbun led its front page with an article headlined “AI Makes New Discoveries in Drug Discovery and Mathematics: Rapid Progress as Reasoning Abilities Improve.” The article reported that cutting-edge AI is beginning to make discoveries in science and mathematics that had previously eluded humans. As scientific discovery becomes increasingly automated and AI generates vast numbers of hypotheses, scientists’ roles are expected to shift from “generating hypotheses” to “designing and governing the process of testing them.”
I asked generative AI to conduct an in-depth investigation into this topic and am sharing the findings for your reference. Please note, however, that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect the actual situation, 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.
0 Comments

サマリア、特許検索競技大会の過去問で「再現率100%」

12/9/2026

0 Comments

 
パテント・インテグレーションは2026年9月10日、生成AIによる特許調査サービス「サマリア」が、2024年度特許検索競技大会アドバンストコースの過去問検証において、電気・機械・化学/医薬の3分野すべてで「再現率100%(上位30件)」を達成したと発表しました。特許調査への生成AI活用が進むなか、検索式の語彙問題という課題を、「類語シソーラス拡張」機能で対応した成果です。(「類語シソーラス拡張」機能とは、過去の拒絶理由通知書10万件以上から、審査官が本願発明と引用文献それぞれの用語・概念を対応するものと認定した事例を抽出し、調査時の類語辞書として整備した機能。)
この発表について、ClaudeとChatGPTに深掘り調査・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。
 
生成AIによる特許調査サービス「サマリア」、特許検索競技大会の過去問検証で全3分野の再現率100%を達成
https://note.com/patent_i/n/nc5d61858bb7d
 
 
Summaria Achieves “100% Recall” on Past Patent Search Competition Questions
On September 10, 2026, Patent Integration announced that its generative AI-powered patent search service, Summaria, had achieved “100% recall (top 30 results)” across all three technical fields—electrical engineering, mechanical engineering, and chemistry/pharmaceuticals—in an evaluation using past questions from the Advanced Course of the 2024 Patent Search Competition. As the use of generative AI in patent searches continues to expand, this achievement demonstrates how the service’s “Synonym Thesaurus Expansion” feature addresses vocabulary-related challenges in formulating search queries.
(The “Synonym Thesaurus Expansion” feature draws on more than 100,000 past notices of reasons for refusal to extract instances in which examiners identified correspondences between terms and concepts used in the inventions under examination and those in cited references. These correspondences are compiled into a synonym dictionary for use in patent searches.)
I asked Claude and ChatGPT to conduct in-depth research and analysis of this announcement, and their findings are presented for your reference. Please note, however, that these AI-generated research and analysis results 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.
0 Comments

世界イノベーションクラスターで東京・横浜は2位

12/9/2026

0 Comments

 
世界知的所有権機関(WIPO)は2026年9月8日、スイス・ジュネーブの本部で「2026世界イノベーション・インデックス(GII)」のイノベーションクラスター・トップ100を発表しました。クラスターの規模ランキングを見ると、中国の「深セン-香港-広州」が2年連続でトップ、2位は「東京-横浜」でした。
このイノベーションクラスター・トップ100について、生成AIに深掘り調査させましたので、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。
 
WIPO、世界のイノベーションクラスター上位100を先行公表、東京・横浜は2位
https://www.jetro.go.jp/biznews/2026/09/8880f2bc1a2ca364.html
 
 
Tokyo–Yokohama Ranks Second Among the World’s Innovation Clusters
On September 8, 2026, the World Intellectual Property Organization (WIPO) announced the top 100 innovation clusters in its Global Innovation Index (GII) 2026 at its headquarters in Geneva, Switzerland. In the ranking by cluster size, China’s Shenzhen–Hong Kong–Guangzhou cluster took the top spot for the second consecutive year, followed by Tokyo–Yokohama in second place.
I asked generative AI to conduct an in-depth study of these top 100 innovation clusters. Please review the findings, bearing in mind that the AI-generated research and analysis are based solely on publicly available information, do not necessarily reflect the actual situation, 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.
0 Comments

Sakana AIの「Fugu Max」「Fugu Ultra v2」

12/9/2026

0 Comments

 
Sakana AIが2026年9月11日に発表したAIオーケストレーター、「Fugu Max」と「Fugu Ultra v2」は、単体で機能するのではなく、複数の外部モデルや専門モデルを動的に選択・統合することで、コスト効率と高度な推論性能の両立を目指す独自の設計思想に基づいています。
発表によれば、Fugu Ultra v2は、「複雑で複数のステップが必要なタスクで、最高の品質を出すこと」を追求した最上位モデルの最新版です。ソフトウェア開発の性能を測る「DeepSWE」や、視覚的な推論・データ解釈を行う「Chartography」などのテストにおいて、他社の「GPT-6 Astra」や「Claude Fable 5.1 / 5」を超える、または同等の世界トップクラスのスコアを叩き出しており、Fugu Maxは、「可能な限り低いコストで、高い成果を出すこと」を突き詰めたモデルで、競合となる「GPT-5.6 Terra」や「Claude Sonnet 5」などと比べて40〜60%も安い価格で、コストを抑えつつも、複数の主要ベンチマーク(SWEFish、Terminal Bench 2.1など)で最高スコアを達成しています。
「Fugu Max」と「Fugu Ultra v2」について、生成AIに深掘り調査させましたので、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。
 
Introducing Fugu Max and Fugu Ultra v2: Orchestrating the Pareto Frontier
September 11, 2026
https://sakana.ai/fugu-max-release/
 
 
Sakana AI’s “Fugu Max” and “Fugu Ultra v2”
“Fugu Max” and “Fugu Ultra v2,” the AI orchestrators announced by Sakana AI on September 11, 2026, are built around a distinctive design philosophy. Rather than operating as standalone models, they dynamically select and integrate multiple external and specialized models, aiming to combine cost efficiency with advanced reasoning capabilities.
According to the announcement, Fugu Ultra v2 is the latest version of the company’s flagship model, designed to “deliver the highest possible quality on complex, multi-step tasks.” On benchmarks such as DeepSWE, which measures software development performance, and Chartography, which assesses visual reasoning and data interpretation, it has achieved world-class scores that match or exceed those of competing models such as GPT-6 Astra and Claude Fable 5.1 / 5.
Fugu Max, meanwhile, is designed to “deliver strong results at the lowest possible cost.” Priced 40–60% lower than competing models such as GPT-5.6 Terra and Claude Sonnet 5, it has achieved the highest scores on several major benchmarks, including SWEFish and Terminal Bench 2.1, while keeping costs down.
I asked generative AI to conduct in-depth research on Fugu Max and Fugu Ultra v2 and am sharing the findings for your reference. Please note that these AI-generated research and analysis results are based solely on publicly available information, may not necessarily reflect the actual situation, 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.
0 Comments

Sony AIが開発した科学的発見予測システム「Hakken」

11/9/2026

0 Comments

 
Sony AIらのチームが2026年9月3日に査読前論文を公開し、時系列知識グラフと大規模言語モデルを統合して未記録の科学的関係を予測するシステム「Hakken」を発表しました。
「Hakken」は過去の論文から知識グラフを構築し、老化関連で150万件の仮説を生成。検証でTP53とBAMBI、RAF1とTNFの新相互作用が2件確認されたということです。
このニュースを生成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。
 
Sony AIのHakkenが老化関連の未知遺伝子関係を発見
https://x.com/i/trending/2097223841047105797
 
 
Hakken, a Scientific Discovery Prediction System Developed by Sony AI
A team including researchers from Sony AI released a preprint on September 3, 2026, introducing “Hakken,” a system that integrates temporal knowledge graphs with large language models to predict previously undocumented scientific relationships.
Hakken constructs knowledge graphs from previously published research papers and generated 1.5 million hypotheses related to aging. According to the report, validation confirmed two novel interactions: one between TP53 and BAMBI, and another between RAF1 and TNF.
I asked generative AI to investigate this news in depth and invite you to review its findings. Please note, however, that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect the actual situation, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

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

低コストでフラッグシップ超え「DeepSeek-V4.1-Flash」

11/9/2026

0 Comments

 
DeepSeekは2026年9月10日、画像も読める新型AIモデル「DeepSeek-V4.1-Flash」を発表しました。DeepSeek-V4.1-Flashは、新世代アーキテクチャの上に構築されたマルチモーダルのMoE(Mixture-of-Experts)型AIモデルで、OpenAIの「GPT-5.6 Sol」、米Anthropicの「Claude Opus 5」などを一部超える性能をより低コストで実現したとしています。
DeepSeek-V4.1-Flashについて、生成AIに深掘り調査させましたので、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。
 
Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.
https://www.deepseek.com/en/news/deepseek-v4-1-flash/
 
 
DeepSeek-V4.1-Flash: Surpassing Flagship Models at a Lower Cost
On September 10, 2026, DeepSeek announced “DeepSeek-V4.1-Flash,” a new AI model that can also interpret images. Built on a next-generation architecture, DeepSeek-V4.1-Flash is a multimodal mixture-of-experts (MoE) model. According to DeepSeek, it outperforms models such as OpenAI’s “GPT-5.6 Sol” and U.S.-based Anthropic’s “Claude Opus 5” in certain areas while delivering this performance at a lower cost.
I asked generative AI to conduct an in-depth investigation into DeepSeek-V4.1-Flash and invite you to review the findings. Please note that the AI-generated research and analysis are based solely on publicly available information, may not necessarily reflect the actual state of affairs, and may contain inaccuracies. Please keep these limitations in mind when reviewing the findings.

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