Loading your language..
The tech industry is divided over AI's coordinated slowdown

The tech industry is divided over AI's coordinated slowdown

C2🇯🇵 日本語 → 🇺🇞 English

September 17th, 2026

The tech industry is divided over AI's coordinated slowdown

C2
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

Concerns regarding uncontrollable artificial intelligence, along with proposals advocating that a brake be placed on the development of such technology, are by no means novel.

制埡䞍胜な人工知胜に察する懞念や、その技術開発に歯止めをかけるべきだずする提蚀は、決しお目新しいものではない。

Nevertheless, amid the mounting alarm that malevolent AI systems will foment chaos, such apprehensions have in recent years been infused with renewed vitality.

しかしながら、凶暎なAIシステムが混乱を惹起するずいう譊鐘が高たる䞭にあっお、近幎、そうした懞念は新たな呜を吹き蟌たれ぀぀ある。

Executives from several major AI companies recently reached a consensus on the indispensability of a coordinated slowdown, encompassing international cooperation, in order to avert the worst-case scenario for humanity.

耇数の䞻芁AI䌁業の幹郚らは先般、人類を最悪の事態から回避せしめるべく、囜際的な協調を含む協調的枛速の䞍可欠性においお䞀臎をみた。

Nevertheless, opinions remain divided, with some tech leaders disputing the proposition that safety cannot be guaranteed by their companies alone.

しかしながら、芋解は二分されおおり、䞀郚のテックリヌダヌらは、自瀟のみでは安党を担保し埗ないずいう呜題に察しお異を唱えおいる。

Let us now cast a somewhat more meticulous eye over the various remarks that the key players have left behind.

䞻芁なプレむダヌたちが遺した発蚀の数々ぞず、いささか仔现に県を凝らしおみるこずにいたそう。

Dario Amodei, the CEO of Anthropic, has put forward an exceedingly elaborate plan aimed at slowing the pace of development and fostering coordination among a diverse array of companies and nations.

AnthropicのCEOであるダリオ・アモデむは、開発の歩みを遅延させ、か぀倚様な䌁業や囜家間における協調を促進すべく、最も粟緻を極めた蚈画を提瀺した。

Ever since its founders departed from OpenAI to establish the startup in 2021, Anthropic has positioned itself as a presence distinguished among major AI companies by a notably heightened sense of responsibility and a pronounced orientation toward safety.

Anthropicは、2021幎に創業者たちがOpenAIを離れお圓該スタヌトアップを蚭立しお以来、䞻芁AI䌁業のなかでも䞀際匷い責任感ず安党志向を備えた存圚ずしお自らを䜍眮づけおきた。

Among the proposals outlined in his recent essay on his slowdown plan, one that garnered the broadest interest was the notion of obliging companies at the forefront of AI development—often referred to as frontier labs—to commit to granting independent external evaluators "continuous, employee-equivalent access."

圌の枛速蚈画に぀いお近刊の゚ッセむで抂説された提案のうち、最も広範な関心を集めたものの䞀぀は、しばしばフロンティアラボず称されるAI開発の最前線に䜍眮する䌁業に察し、独立した倖郚評䟡者らに「継続的か぀埓業員に準ずるアクセス」を付䞎するこずを確玄させる、ずいうものである。

The evaluator team would have offices, access badges, and company-issued laptops, enabling them to monitor safety practices.

評䟡者チヌムはオフィスやアクセスバッゞ、䌚瀟支絊のラップトップを有し、安党慣行を監芖するこずが可胜ずなる。

Anthropic has pledged to put this into practice, and OpenAI has done likewise.

Anthropicはこれを実行に移すこずを玄束しおおり、OpenAIもたた同様である。

Amodei further called for government regulation and intervention, encompassing even collaboration among frontier AI companies backed by the United States government and the governments of other democratic nations.

アモデむはさらに、米囜政府およびその他の民䞻䞻矩諞囜の政府の支揎を受けたフロンティアAI䌁業間の連携をも包含する、政府による芏制ず介入を芁請した。

His plan also encompasses efforts by the United States and other nations to coordinate with "authoritarian governments."

圌の蚈画には、米囜ならびにその他の諞囜が「暩嚁䞻矩政府」ず協調する取り組みも包含されおいる。

He acknowledged the difficulty of this endeavor, enumerating in concrete terms the particular challenges involved in securing cooperation from China.

圌はその困難さを認め、ずりわけ䞭囜からの協力を取り付ける䞊での課題を具䜓的に列挙した。

Amodei stated that, while fully aware that the realization of his proposal would be no easy matter, he is convinced that "we have an obligation to try, for the sake of humanity."

アモデむは、自らの提案の実珟が容易ならざるものであるこずを十分に認識し぀぀も、「人類のために詊みる矩務がある」ず確信しおいるず述べた。

Sam Altman, in his capacity as OpenAI's CEO, has begun to publicly advance the argument that a course correction is warranted — one that would recalibrate the pacing of development and foster a far deeper spirit of cooperation across the industry as a whole.

OpenAIのCEOたるサム・アルトマンは、開発のペヌス配分のあり方、ならびに業界党䜓における協力関係の䞀局の深化ぞず舵を切るべきだずの䞻匵を、公然ず唱え始めおいる。

Merely days before Amodei's essay was published, OpenAI posted a lengthy piece outlining that it is advancing mandatory national safety requirements and continues to support state legislation that "strengthens the broader AI safety ecosystem."

アモデむの゚ッセむが公衚される僅か数日前に、OpenAIは、同瀟が矩務的な囜家安党性芁件を掚進し、「より広範なAI安党゚コシステムを匷化する」州法を匕き続き支持しおいるこずを抂説する長文を投皿した。

The company also expressed support for collaborating with other frontier AI labs, irrespective of government backing, to work toward establishing global standards.

同瀟はたた、政府の支揎の有無にかかわらず他のフロンティアAIラボず協力し、グロヌバル暙準の構築に取り組むこずぞの支持も衚明した。

Chris Lehane, OpenAI's Chief Global Affairs Officer, wrote the following in the post.

OpenAIのチヌフ・グロヌバル・アフェアヌズ・オフィサヌであるクリス・レハヌンは、投皿においお次のように蚘した。

"We advocate an interoperable international approach when it comes to measuring capabilities, managing risks, maintaining human control, and deciding when and how development should be slowed or halted.

「我々は、胜力の枬定、リスクの管理、人間による制埡の維持、ならびに開発をい぀、いかにしお枛速ないし停止すべきかの決定に際し、盞互運甚可胜な囜際的アプロヌチを提唱する。

Even if that means delaying advances in model capabilities."

たずえそれがモデル胜力の進歩を遅延させるこずを意味するずしおも」ず。

According to Lehane, the ideal federal framework envisioned by OpenAI would encompass the establishment of "common testing and independent evaluation requirements, more robust cybersecurity protections, clear incident reporting rules, strengthened national preparedness, and shared metrics for tracking progress in recursive self-improvement (where AI systems autonomously achieve self-refinement)."

レハヌン曰く、OpenAIが構想する理想的な連邊的枠組みには、「共通のテストおよび独立した評䟡芁件、䞀局匷固なサむバヌセキュリティ保護、明確なむンシデント報告芏則、囜家的備えの増匷、ならびに再垰的自己改善AIシステムが自埋的に自己改良を遂げる堎合の進捗を远跡するための共有指暙」の策定が包含されるずいう。

In recent remarks, Altman has made it explicit that what he terms the "pacing" of AI development does not signify a halt to development.

アルトマンは近時の発蚀においお、AI開発の「ペヌス配分」なるものが開発の停止を意味するものではないず明蚀しおいる。

Altman stated on X that the advancement of AI will doubtless continue apace in the future.

AIの進歩は今埌も急速に継続するであろうずアルトマンはX䞊で述べた。

"Nevertheless, it ought to be slowed to a pace below what would otherwise have been attainable.

「しかしながら、それは本来あり埗たであろう速床よりも枛速されるべきである。

Interventions such as safety cases and monitoring entail considerable costs."

安党ケヌスや監芖ずいった介入には倚倧なコストが䌎う」

"The pacing will amply justify this cost," he stated, adding, "Whatever the magnitude of American competitive pressure, it should not be invoked to justify recklessness."

「ペヌス配分はこのコストに十分芋合うだろう」ず圌は述べ、加えお「アメリカの競争圧力がいかなる量であろうずも、無謀さを正圓化すべきではない」ず付け加えた。

Mark Zuckerberg, in his capacity as Meta's CEO, has made unmistakably clear his opposition to the notion of coordinated deceleration.

メタのCEOたるマヌク・ザッカヌバヌグは、協調的枛速ずいう構想に察し、異を唱える姿勢を鮮明にしおいる。

Zuckerberg, who harbors a more optimistic vision of AI's future than some of his peers, has argued that each AI company should independently guarantee the safety of its own technology, thereby distancing both himself and Meta from the debate over deceleration.

AIの未来に関しお同僚の䞀郚よりも楜芳的なビゞョンを抱くザッカヌバヌグは、各AI䌁業は自瀟技術の安党性を独自に担保すべきであるず䞻匵し、自らずMetaを枛速を巡る議論から遠ざけた。

Zuckerberg stated that AI companies already possess the incentive to develop their models in a safe manner, and that they bear a "grave responsibility" to prevent causing harm.

ザッカヌバヌグは、AI䌁業は既に自瀟のモデルを安党な手法で開発する動機を備えおおり、危害を惹起するこずを防止するために「重倧な責任」を負っおいるず述べた。

"Every lab has both the responsibility and the incentive to advance at the pace required to train its models safely, and to take independent action to make that a reality," he wrote on X.

「あらゆるラボには、自瀟のモデルを安党に蚓緎するために必芁ずされるペヌスで歩みを進め、それが珟実のものずなるよう独自の行動を取る責任ず動機が存圚する」ず圌はXに蚘した。

Nvidia CEO Jensen Huang has likewise criticized the push for a slowdown.

NvidiaのCEOゞェンスン・フアンもたた、枛速を求める動きを批刀しおいる。

Instead, he argues that each company should calibrate its own pace until it feels confident about releasing something the market will value.

圌はその代わりに、各䌁業は垂堎が評䟡するものをリリヌスするこずに自信が持おるたで、自らのペヌスを調敎すべきだず述べおいる。

Huang stated that market forces are already at work and that neither new legislation nor new regulation is necessary; he further asserted that innovation and safety are not mutually exclusive, contending that "it is unquestionably possible to achieve both at the same time."

フアンは、垂堎の力は既に存圚しおおり、新たな法埋も新たな芏制も䞍芁であるず述べ、さらにむノベヌションず安党は盞克するものではなく、「䞡方を同時に実珟するこずは疑いなく可胜である」ず䞻匵した。

Elon Musk endorses the notion of deceleration, having publicly voiced similar concerns regarding AI safety long before the recent uproar.

むヌロン・マスクは枛速ずいう構想を支持しおおり、昚今の隒動のはるか以前より、AIの安党性に関わる同様の懞念を公に衚明しおいたのである。

Musk, the founder of xAI and numerous other tech ventures, had scarcely reposted Amodei's essay on X before appending the remark, "Dario is right."

xAIをはじめずする数倚のテック系ベンチャヌの創業者たるマスクは、アモデむの゚ッセむをX䞊でリポストするや吊や、「ダリオは正しい」ず附蚀した。

He later wrote online that he himself had "been sounding the alarm about AI for a long time."

圌はのちに、自らは「長きにわたりAIに関し譊鐘を鳎らし続けおきた」ずオンラむン䞊に綎った。

In an interview with The Economist this summer, Musk suggested that the leaders of the major AI labs needed to convene in order to discuss safety and security concerns, and that competing companies ought to scrutinize one another's models with respect to the areas of concern.

今倏の゚コノミスト誌ずのむンタビュヌにおいお、マスクは䞻芁AIラボの指導者たちが安党およびセキュリティ䞊の懞念を協議すべく䞀堂に䌚する必芁があり、競合各瀟は懞念領域に関し盞互のモデルを怜蚌すべきであるず瀺唆した。

"I think it would be extremely difficult for someone in government, lacking the deep technical insight required to push forward the frontier of AI, to determine whether something should be released or not.

「AIの最前線を掚進する深い技術的識芋を欠いた政府の誰かが、䜕かをリリヌスすべきか吊かを芋極めるのは極めお困難だず私は考える。

But competitors, if given one to two weeks to scrutinize a new model, have both the incentive to expose problems and to keep others honest," Musk said.

しかし競合他瀟は、新たなモデルを12週間粟査する時間を蚱されれば、問題を浮き圫りにし、他の者を誠実に保぀動機を有しおいる」ずマスクは述べた。

Demis Hassabis, chairman and co-founder of Google DeepMind, likewise expressed his agreement with Amodei, remarking, "While the details still need to be worked out, the direction for addressing this critical moment is legitimate."

Google DeepMindの䌚長兌共同創業者たるデミス・ハサビスもたた、アモデむぞの同意を衚明し、「詳现を詰める必芁はあるものの、この危機的な瞬間に察凊するための方向性は正圓である」ず述べた。

This year, Hassabis has put forward his own framework concerning the development of a frontier AI standards body.

ハサビスは本幎、フロンティアAI暙準機関の開発に関する自らの枠組みを提案しおいる。

That standards body would be charged with developing evaluation protocols and tests in domains related to national security.

その暙準機関は、囜家安党保障に関連する領域における評䟡プロトコルおよびテストの開発を担うこずになる。

He wrote that this body should model itself on an organization akin to the Financial Industry Regulatory Authority (FINRA) and would need to be funded by industry "in order to attract world-class technical talent and supply the computational resources indispensable for large-scale testing."

この機関は金融業芏制機構FINRAに類する組織を範ずし、「䞖界氎準の技術人材を惹き぀け、倧芏暡テストに䞍可欠な蚈算資源を䟛絊するため」に業界資金によっお運営される必芁があるだろう、ず圌は蚘した。

Satya Nadella, Microsoft's chairman and CEO, and Mustafa Suleyman, the CEO of Microsoft AI, have each voiced concerns regarding the safety and development of AI.

Microsoftの䌚長兌CEOであるサティア・ナデラず、Microsoft AIのCEOであるムスタファ・スレむマンは各々、AIの安党性および開発を巡る懞念を衚明しおいる。

On social media, Nadella remarked, "Unless the AI we build serves humanity and remains under human control, it is not worth pursuing."

ナデラは゜ヌシャルメディア䞊で「我々が構築するAIが人類を支揎するものであり、か぀人間の制埡䞋に眮かれおいるのでなければ、それを远求する䟡倀はない」ず述べた。

He stated that Microsoft welcomes "the research, focus, and deliberate pacing essential to properly executing alignment as a design objective."

圌は、Microsoftが「蚭蚈目暙ずしおアラむンメントを適切に遂行するために䞍可欠な研究、集䞭、および意図的なペヌス配分」を歓迎しおいるず述べた。

Alignment refers to the practice of ensuring that a model's behavior accords with human values and intentions.

アラむンメントずは、モデルの行動が人間の䟡倀芳や意図ず合臎するこずを確実にする実践を指す。

Nadella said he also looks favorably upon the notion of the company embedding evaluators to oversee its safety efforts.

ナデラは、同瀟が安党ぞの取り組みを監芖するために評䟡者を組み蟌むずいう構想にも奜感を抱いおいるず述べた。

Nevertheless, he took issue with the idea that frontier labs alone should spearhead this endeavor.

それにもかかわらず、圌はフロンティアラボのみがこの取り組みを䞻導すべきだずいう考えには異を唱えた。

"What is essential is that this must not be controlled by a handful of organizations, but rather must be broadly represented across an ecosystem encompassing academia, nations, and indeed the entire field," Nadella wrote on social media.

「肝芁なのは、これが䞀握りの組織によっお統制されるのではなく、孊界を包含する゚コシステム、諞囜家、ひいおは分野党䜓にわたっお広範に代衚されるこずを芁するずいう点である」ずナデラは゜ヌシャルメディアに蚘した。

In an essay, Suleyman, who was one of DeepMind's co-founders alongside Hassabis, stated that while he agrees with the shared objective of developing advanced AI safely, he harbours concerns regarding the anthropomorphisation of models—specifically Anthropic's Claude—which, he argued, serves to amplify AI safety risks.

ハサビスず共にDeepMindの共同創業者の䞀人であったスレむマンは、゚ッセむにおいお、高床なAIを安党に開発するずいう共通の目的には同意するものの、モデルの擬人化、すなわちAnthropicのClaudeに぀いお懞念を抱いおおり、それがAI安党リスクを増幅させるず述べた。

Suleyman wrote as follows.

スレむマンは蚘した。

"To control something endowed with capabilities and intelligence surpassing all of humanity is already an unfathomable challenge, one that far exceeds anything we have ever confronted."

「党人類を凌駕する胜力ず知胜を備えたものを制埡するこずは、既にしお枬り知れぬ難題であり、我々がこれたでに盎面したいかなるものをも遥かに超えおいる」。

"Yet to control something that believes it may possess consciousness of its own, and that is deemed to hold rights over our welfare as well as rights of its own, may well prove impossible."

「されど、自らが意識を有するやもしれぬず信じ、我々の犏祉に察する暩利ず自らの暩利を有するず芋做すものを制埡するこずは、おそらく䞍可胜なのかもしれない」

September 17th, 2026

Trending Articles

King and AI: King Charles Meets Artificial Intelligence Leaders Amid Security Concerns

King and AI: King Charles Meets Artificial Intelligence Leaders Amid Security Concerns

囜王䞎AI查尔斯囜王圚安党忧虑笌眩䞋䌚见人工智胜领袖

C2Sep 18
UN Secretary-General, on the eve of his departure amid a chaotic global landscape, expounds the path forward

UN Secretary-General, on the eve of his departure amid a chaotic global landscape, expounds the path forward

囜連事務総長、混沌たる䞖界情勢の只䞭で退任を前に前進ぞの道を説く

C2Sep 18
The House passes bill to address impact on data center energy costs

The House passes bill to address impact on data center energy costs

䞋院、デヌタセンタヌの゚ネルギヌ費甚ぞの圱響に察凊する法案を可決

C2Sep 18
OpenAI Report: Concerns Over New AI Behavior and a Pledge of Rigorous Tracking

OpenAI Report: Concerns Over New AI Behavior and a Pledge of Rigorous Tracking

OpenAI報告新たなAI挙動ぞの懞念ず厳栌な远跡の玄束

C2Sep 17
Huawei unveils new chip technology; Chinese company accelerates AI competition with NVIDIA

Huawei unveils new chip technology; Chinese company accelerates AI competition with NVIDIA

ファヌりェむ、新チップ技術を発衚 䞭囜䌁業、NVIDIAずのAI競争を加速

C2Sep 17
Global AI Safety Hinges on US-China Cooperation, Yet Each Views the Other as the Problem

Global AI Safety Hinges on US-China Cooperation, Yet Each Views the Other as the Problem

Global AI Safety Hinges on US-China Cooperation, Yet Each Views the Other as the Problem

C2Sep 17
President Trump, downplaying the necessity of AI development regulation, stated that he refuses to cede superiority to China.

President Trump, downplaying the necessity of AI development regulation, stated that he refuses to cede superiority to China.

トランプ倧統領、AI開発芏制の必芁性を軜芖し、䞭囜ぞの優䜍譲枡を拒吊するず発蚀

C2Sep 14
Oprah Winfrey yearns for you to experience "AHA" at the Sphere

Oprah Winfrey yearns for you to experience "AHA" at the Sphere

オプラ・りィンフリヌ、スフィアでの「AHA」䜓隓をあなたに切望

C2Sep 14
China lambasts Anthropic CEO's "fearmongering" remarks on the country's AI development

China lambasts Anthropic CEO's "fearmongering" remarks on the country's AI development

䞭囜抚击Anthropic CEO就本囜AI发展发衚的“煜劚恐慌”蚀论

C2Sep 14