Loading your language..
Anthropic announced that its own model Claude is assisting in the development of a next-generation version.

Anthropic announced that its own model Claude is assisting in the development of a next-generation version.

C1🇯🇵 日本語 → 🇺🇞 English

September 18th, 2026

Anthropic announced that its own model Claude is assisting in the development of a next-generation version.

C1
Please note: This article has been simplified for language learning purposes. Some context and nuance from the original text may have been modified or removed.

🇺🇞 English

In an announcement on Thursday, the AI research institute stated that Anthropic's Claude is helping the company develop next-generation, more intelligent models.

AnthropicのClaudeは、同瀟が次䞖代の、より知胜の高いモデルを開発する䞊で助力ずなっおいるず、このAI研究所は朚曜日の発衚においお述べた。

Claude leads 26% of Anthropic's model research and development, which, the company said, means it can complete most tasks "end-to-end from a high-level prompt" while still under human supervision.

ClaudeはAnthropicのモデル研究開発の26%を䞻導しおおり、これは人間の監督䞋にありながら、「高レベルなプロンプトから゚ンドツヌ゚ンドで」ほずんどのタスクを完了できるこずを意味するず同瀟は述べた。

That said, the model does not yet operate fully autonomously.

ずはいえ、このモデルは未だ完党に自埋的に動䜜しおいるわけではない。

Nevertheless, roughly 90% of the company's research and development is carried out through "collaboration" with Claude, and Anthropic stated that this means the model can handle "the majority of the work under close human direction."

それにもかかわらず、同瀟の研究開発のおよそ90%はClaudeずの「協働」を通じお遂行されおおり、このこずは、モデルが「人間による綿密な指瀺の䞋で䜜業の倧郚分」を担い埗るこずを意味するず、Anthropicは述べた。

This announcement was made under the leadership of Dario Amodei, Anthropic's CEO, at a time when some leading figures in the AI world are calling for a slowdown in technological development out of safety concerns.

この発衚は、AI界の䞀郚の䞻芁人物が安党䞊の懞念から技術開発の枛速を呌びかけおいる䞭で、AnthropicのCEOであるダリオ・アモデむ氏が䞻導する圢で行われた。

In considering the pace of AI development, the company stated in a blog post that it "should do everything possible to minimize the gap between what frontier labs know and what the general public knows."

AIの開発ペヌスを怜蚎するにあたり、同瀟はブログ投皿で、「フロンティアラボが知っおいるこずず䞀般公䌗が知っおいるこずのギャップを最小限に抑えるために可胜な限りすべおを行うべきだ」ず述べた。

"This means better measuring AI progress, reporting it publicly, and giving society the opportunity to decide how to use this information."

「これは、AIの発展をよりよく枬定し、それを公に報告し、この情報をどのように掻甚するかを瀟䌚が決定する機䌚を䞎えるこずを意味する。」

The company stated that models accelerating their own development could "make it more difficult for humans to understand and control these systems."

同瀟は、モデルが自らの開発を加速させるこずは「人間がこれらのシステムを理解し、制埡するこずをより困難にする」可胜性があるず述べた。

It argued that sharing these metrics could lead to a deeper understanding of how close leading AI labs are to recursive self-improvement—that is, the ability of models to autonomously build their successors.

これらの指暙を共有するこずで、䞻芁なAIラボが再垰的自己改善、すなわちモデルが自埋的に埌継モデルを構築する胜力にどれだけ近づいおいるかに぀いお、より深く理解できるようになる可胜性があるず䞻匵した。

Anthropic also urged other AI developers to regularly share similar metrics, and by encouraging the use of a public methodology, it made it possible to compare the figures over time and potentially across labs.

Anthropicはたた、他のAI開発者に察しおも同様の指暙を定期的に共有するよう促し、公的な方法論の䜿甚を奚励するこずで、数字を経時的に、そしお朜圚的にラボ間で比范できるようにした。

Anthropic's disclosure did not make clear how close the company believes it is to achieving recursive self-improvement, but the pace at which Claude is contributing more and more to research and development is noteworthy.

Anthropicの開瀺からは、同瀟が再垰的自己改善の達成にどれだけ近いず考えおいるかは䞍明だったが、Claudeが研究開発にたすたす貢献するペヌスは泚目に倀する。

The share of work that Claude "leads" — that is, carries out most of while remaining under human supervision — was zero in February.

Claudeが「䞻導」する、぀たり人間の監督䞋にありながら倧郚分を行う䜜業の割合は、2月にはれロだった。

Six months later, it was leading a quarter of research and development work, reaching that threshold in August.

6か月埌には研究開発䜜業の4分の1を䞻導し、8月にその基準に達した。

The company further shared details of the agent oversight measures it has implemented, noting that as of August, roughly 30,000 agents were engaged in research and engineering work.

同瀟はさらに、導入枈みの゚ヌゞェント監督措眮の詳现を共有し、8月時点で玄3䞇の゚ヌゞェントが研究および゚ンゞニアリング業務に埓事しおいるず指摘した。

According to the company, such oversight measures are crucial for verifying how frequently misconduct by agents is detected by monitoring systems.

同瀟によれば、こうした監督措眮は、゚ヌゞェントによる䞍正行為が監芖システムによっおどの皋床の頻床で怜出されるかを怜蚌する䞊で極めお重芁であるずいう。

Anthropic recently committed to stationing external third-party evaluators on-site within the company to monitor its safety efforts.

Anthropicは最近、安党ぞの取り組みを監芖するため、瀟内に垞駐する倖郚の第䞉者評䟡者を蚭眮するこずを玄束した。

Last week, a researcher at Anthropic resigned while leaving behind a grave warning about the threat this technology poses to humanity, and this became the catalyst for much of the recent debate surrounding AI safety.

先週、Anthropicの研究者が、この技術が人類に及がす脅嚁に぀いお深刻な譊告を残しお蟞任したこずが、AI安党性をめぐる最近の議論の倚くが始たる契機ずなった。

Since then, Sam Altman, the CEO of OpenAI, Elon Musk, and other tech leaders have endorsed the idea of slowing down development, while other tech leaders and President Donald Trump are opposed to it.

OpenAIのCEOであるサム・アルトマン氏、むヌロン・マスク氏、その他の技術指導者たちはその埌、開発の枛速ずいう考えを支持しおいるが、他の技術指導者たちやドナルド・トランプ倧統領は反察しおいる。

September 18th, 2026

Trending Articles

Warren Buffett to step down as Berkshire Hathaway chairman, next step in succession plan

Warren Buffett to step down as Berkshire Hathaway chairman, next step in succession plan

Warren Buffett to step down as Berkshire Hathaway chairman, next step in succession plan

C1Sep 18
President Trump says he will ban CNN, MS NOW, and Politico from the White House, targeting further news organizations

President Trump says he will ban CNN, MS NOW, and Politico from the White House, targeting further news organizations

トランプ倧統領、CNN、MS NOW、ポリティコのホワむトハりス立ち入りを犁止するず発蚀、さらなる報道機関を暙的に

C1Sep 18
The Bank of Japan raises its policy interest rate to 1.25%, the highest level in 31 years.

The Bank of Japan raises its policy interest rate to 1.25%, the highest level in 31 years.

日銀、政策金利を1.25%に匕き䞊げ、31幎ぶりの高氎準に

C1Sep 18
The reasons and implications behind the Fed's decision to raise interest rates now

The reasons and implications behind the Fed's decision to raise interest rates now

FRBが今、利䞊げに螏み切る理由ずその意味

C1Sep 17
US stocks post biggest gain in six weeks amid falling oil prices and declining bond yields

US stocks post biggest gain in six weeks amid falling oil prices and declining bond yields

米囜株、原油安ず債刞利回り䜎䞋で6週間ぶり最倧の䞊昇

C1Sep 17
FRB raises key interest rate for first time in 3 years, defying Trump's demand for cuts

FRB raises key interest rate for first time in 3 years, defying Trump's demand for cuts

FRB raises key interest rate for first time in 3 years, defying Trump's demand for cuts

C1Sep 17
Unemployment insurance claims fell to 196,000, the lowest level since mid-July—layoffs remain subdued

Unemployment insurance claims fell to 196,000, the lowest level since mid-July—layoffs remain subdued

倱業保険申請件数は19侇6000件に枛少、7月䞭旬以来の䜎氎準—解雇は䟝然ずしお䜎調

C1Sep 17
President Trump dismisses AI risks as a "hoax," asserting a "pathological conspiracy" surrounding AI and data centers

President Trump dismisses AI risks as a "hoax," asserting a "pathological conspiracy" surrounding AI and data centers

トランプ倧統領、AIのリスクを「デマ」ず䞀蹎し、AIずデヌタセンタヌを巡る「病的な陰謀」を䞻匵

C1Sep 14
Supreme Court to Decide Whether to Hear Maine Lobstermen's Challenge to GPS Boat-Tracking Requirement

Supreme Court to Decide Whether to Hear Maine Lobstermen's Challenge to GPS Boat-Tracking Requirement

Supreme Court to Decide Whether to Hear Maine Lobstermen's Challenge to GPS Boat-Tracking Requirement

C1Sep 14