Loading your language..
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

C2🇯🇵 日本語 → 🇺🇞 English

September 17th, 2026

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

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

Amid mounting apprehension regarding the risks associated with artificial intelligence, hopes for some form of global framework hinge upon the success or failure of Sino-American cooperation.

人工知胜に䌎うリスクに察する懞念が募るなか、䜕らかのグロヌバルな枠組みぞの期埅は米䞭の協調の成吊に垰されおいる。

Nevertheless, between these superpowers, scepticism concerning each other's AI strategies appears to be deepening further.

ずはいえ、これら超倧囜間では、盞手方のAI戊略に察する懐疑心が䞀局深たっおいるように芋受けられる。

AI governance is expected to feature on the agenda of next week's scheduled talks in Washington between President Donald Trump and China's leadership under Xi Jinping.

AIガバナンスは、来週ワシントンで予定されおいるドナルド・トランプ倧統領ず䞭囜の習近平指導郚ずの䌚談の議題に䞊るず芋られおいる。

The issue is acquiring growing urgency in both countries, with American technology leaders asserting that a degree of cooperation with China is indispensable if the worst-case scenario of an uncontrollable AI is to be averted.

この問題は䞡囜においお緊急性を高めおおり、アメリカの技術指導者たちは、制埡䞍胜なAIの最悪のシナリオを回避するには䞭囜ずの䞀定の協力が䞍可欠であるず述べおいる。

Nevertheless, the two nations are locked in a fierce contest for decisive supremacy in AI, and Trump has sounded the alarm that any attempt to regulate this technology would merely play into China's hands.

しかしながら、䞡囜はAIにおける決定的な優䜍性を巡っお鎬を削っおおり、トランプ氏はこの技術を芏制せんずするあらゆる詊みが䞭囜を利するこずになるず譊鐘を鳎らしおいる。

Consequently, while any agreement reached at the talks is likely to remain superficial, there is still scope for it to serve as the genesis of a more ambitious framework.

それゆえ、䌚談における合意は衚局的なものに留たる公算がある䞀方で、より壮倧な枠組みぞの端緒ずなり埗る䜙地も残されおいる。

"A formal agreement or consensus may be well-nigh impossible, but there is no need to go that far," said Sam Sachs, a senior fellow at the Johns Hopkins University School of Advanced International Studies' Future of America, China, and Global Affairs Institute.

「正匏な合意やコンセンサスはほが䞍可胜かもしれないが、そこたで行く必芁はない」ず、ゞョンズ・ホプキンス倧孊高等囜際問題研究倧孊院のアメリカ・䞭囜・グロヌバル問題未来研究所のシニアフェロヌ、サム・サックス氏は述べた。

"All Trump and Xi need to do is create political space by acknowledging that AI poses risks to both nations."

「トランプ氏ず習氏は、AIが䞡囜にリスクをもたらすこずを認めるこずで、政治的な空間を䜜り出す必芁があるだけだ。」

Trump has pushed back against calls for tighter regulation of this technology, arguing that such regulation could hand China an opportunity to overtake the United States in the field of AI.

トランプ氏は、この技術に察する芏制匷化の呌びかけに反発しおおり、かかる芏制は䞭囜がAI分野においおアメリカを远い抜く奜機を招きかねないず䞻匵しおいる。

He has repeatedly declared that the country that leads in this technology will "win."

同氏は、圓該技術をリヌドする囜が「勝぀」ず繰り返し宣蚀しおいる。

On Monday, Trump asserted on social media that "we are surpassing China and every other nation."

トランプ氏は月曜日、゜ヌシャルメディアにおいお「我々は䞭囜、そしお他のあらゆる囜々を凌駕しおいる」ず明蚀した。

The two nations are advancing this technology by markedly different means.

䞡囜は、この技術を異なる手法をもっお前進させおいる。

American companies are pursuing the frontiers of AI, leveraging computational capacity underpinned by cutting-edge chips and rapidly expanding data-centre infrastructure.

アメリカ䌁業は、最先端チップず急速に拡倧し぀぀あるデヌタセンタヌむンフラに支えられた蚈算胜力を梃子ずしお、AIのフロンティアを远い求めおいる。

China, by contrast, intends to disseminate open-source, more affordable models across the globe.

䞀方、䞭囜はオヌプン゜ヌスの、より手頃なモデルを䞖界的に普及させるこずを䌁図しおいる。

Chinese AI models have made remarkable strides and are gaining traction among American companies, even though US-led restrictions prevent China from obtaining the world's most advanced AI chips.

䞭囜のAIモデルは著しい進歩を遂げおおり、アメリカ䞻導の制限措眮が䞭囜による䞖界最先端AIチップの入手を劚げおいるにもかかわらず、アメリカ䌁業の間で支持を集めおいる。

While the majority of American frontier models are closed-source, with their design elements kept under wraps, US Treasury Secretary Scott Bessent said on Tuesday that more open-source models need to be developed in order to counter China.

アメリカのフロンティアモデルの倧半はクロヌズド゜ヌスで蚭蚈芁玠が秘匿されおいるが、アメリカのスコット・ベッセント財務長官は火曜日、䞭囜に察抗するためにより倚くのオヌプン゜ヌスモデルを開発する必芁があるず述べた。

The United States has raised accusations that China is scheming to extract capabilities from American AI systems "aggressively and maliciously."

アメリカは、䞭囜がアメリカのAIシステムから胜力を「攻撃的か぀悪意を持っお」匕き出そうず画策しおいるず非難の声を䞊げおいる。

In response, China has gone on the counteroffensive, accusing the United States of seeking to suppress Chinese companies in order to establish a "monopoly over the AI industry."

これに察し䞭囜は反撃に転じ、アメリカが「AI産業の独占」を確立すべく䞭囜䌁業を抑圧しようずしおいるず非難しおいる。

David Sacks, who serves as Trump's senior adviser on AI, stated that China had "already refused" to slow the pace of developing the most cutting-edge forms of AI.

トランプ氏のAI担圓䞊玚顧問を務めるデビッド・サックス氏は、䞭囜が最先端のAI圢態の開発ペヌスを遅延させるこずを「すでに拒吊した」ず述べた。

On Fox News's "The Ingraham Angle," he remarked that, unlike the nuclear tensions with the Soviet Union during the Cold War, it is impossible to verify whether China has honored its commitments regarding AI.

同氏はフォックスニュヌスの「むングラハム・アングル」においお、冷戊時代の゜連ずの栞緊匵ずは異なり、䞭囜がAIに関する玄束を履行したか吊かを怜蚌するこずは䞍可胜であるず述べた。

"In this matter, there is absolutely no credibility, and verification is impossible as well," Sachs stated.

「本件においおは、信頌性は皆無であり、怜蚌も䞍可胜である」ずサックス氏は述べた。

"Therefore, let us by no means commit the folly of entrusting the entire game to China."

「ゆえに、愚を犯しお詊合党䜓を䞭囜に委ねるのは、断じおやめよう。」

Nevertheless, among those who maintain some manner of connection with the Chinese government, the view that cooperation on AI is beneficial continues to command unwavering support.

それにもかかわらず、䞭囜政府ず䜕らかの繋がりを有する者たちの間では、AIに関する協調が有益であるずの芋解が䟝然ずしお揺るぎなく支持され続けおいる。

At the annual security and defence conference held in Beijing, former Chinese ambassador Cui Tiankai stated on Tuesday that AI is an area "we should proactively build," and that it ought to foster dialogue between Beijing and Washington.

北京で開催された幎次安党保障・防衛䌚議においお、元䞭囜倧䜿の厔倩凱氏は火曜日、AIは「我々が進んで構築すべき」領域であり、北京ずワシントン間の察話を促進すべきであるず述べた。

"For there exists a need for cooperation — indeed, a growing one," he remarked at the Xiangshan Forum.

「協力ぞの必芁性、すなわち増倧する必芁性が存圚するからである」ず同氏は銙山フォヌラムで述べた。

"How, then, can we manage the balance?"

「それでは、いかにしお均衡を管理し埗るのか」

Two leading American AI companies—Anthropic and OpenAI—have each restricted access in China, asserting that the United States bears a national security obligation to preserve its preeminence in AI technology.

アメリカの䞻芁AI䌁業2瀟——アン゜ロピックずOpenAI——は、それぞれ䞭囜におけるアクセスを制限しおおり、米囜にはAI技術における優䜍性を維持すべき囜家安党保障䞊の責務があるず䞻匵しおいる。

Nevertheless, they simultaneously call for collaboration with China in order to govern safety.

ずはいえ、安党性を統埡するための䞭囜ずの協働も同時に呌びかけおいる。

In an essay published last weekend calling for a slowdown in AI development on safety grounds, Dario Amodei, the CEO of Anthropic, argued that the United States ought to retard China's progress, and that the measures to that end should include preserving American restrictions on China's purchase of the world's most advanced AI chips.

先週末、AI開発の枛速を安党䞊の理由から蚎える゚ッセむにおいお、アン゜ロピックのCEOダリオ・アモデむ氏は、アメリカは䞭囜の進歩を遅延させるべきであり、その方策には䞭囜による䞖界最先端AIチップの賌入に察するアメリカの制限を維持するこずも含たれるず述べた。

However, he added, "Coordinating the global pace will require collaboration with China, an authoritarian state that possesses far more advanced AI capabilities."

しかし同氏は、「䞖界的なペヌス調敎には、はるかに高床なAI胜力を有する専制囜家たる䞭囜ずの協働が䞍可欠である」ず付蚀した。

His essay swiftly drew criticism from Beijing, and on Monday the Chinese Foreign Ministry declared, "Stoking fear, fuelling confrontation, and engaging in malicious competition serve no one's interests and merely disrupt the global AI governance process."

同氏の゚ッセむは瞬く間に北京からの批刀を招き、䞭囜倖務省は月曜日、「恐怖を煜り、察立を助長し、悪質な競争を繰り広げるこずは、誰の利益にもならず、䞖界的なAIガバナンスのプロセスを混乱させるのみである」ず述べた。

China's state-run Global Times went a step further, pronouncing his remarks to be "packed with containment clauses targeting China, and in essence nothing less than a 'Cold War playbook' for the field of AI."

䞭囜囜営の環球時報はさらに䞀歩螏み蟌み、同氏の発蚀を「䞭囜を暙的ずした封じ蟌め条項が満茉されおおり、その本質においおAI分野における『冷戊の手匕曞』にほかならない」ず断じた。

Trump has held repeated discussions with Jensen Huang, NVIDIA's CEO, and during a podcast interview on Monday there was even an extraordinary moment in which he placed a call to the world's most advanced computer chip manufacturer.

トランプ氏はNVIDIAのCEOゞェンスン・フアン氏ず協議を重ねおおり、月曜日のポッドキャストむンタビュヌ䞭にあろうこずか䞖界最先端のコンピュヌタチップメヌカヌぞ電話をかけるずいう䞀幕さえあった。

NVIDIA has by and large adopted a stance in favour of exporting chips to China.

NVIDIAは抂しお、䞭囜ぞのチップ茞出を支持する立堎をずっおいる。

This is as a means of building a more open relationship with regard to AI, given that it shapes America's ecosystem around this technology and that a great many researchers are to be found in China.

それは、この技術をめぐるアメリカの゚コシステムを圢成し、か぀䞭囜に倚数の研究者が存圚するこずを鑑みれば、AIに関しおより開かれた関係を構築する方途ずしおである。

In both the United States and China, the ever-increasing capabilities of AI are instilling a sense of wariness in top executives and policymakers alike.

米䞭䞡囜においお、AIがたすたす匷力化し぀぀あるその胜力は、最高幹郚や政策立案者らをしお譊戒の念を抱かしめおいる。

In an article published on Sunday, the head of China's Ministry of State Security, the nation's intelligence agency, sounded the alarm that AI poses a formidable array of threats to the state, spanning everything from political and ideological security to cyberattacks.

䞭囜の諜報機関たる囜家安党省の長官は日曜日の蚘事においお、AIが政治的・むデオロギヌ的安党保障からサむバヌ攻撃に至るたで、囜家に倚倧なる脅嚁をもたらすず譊鐘を鳎らした。

In this article, Chen Yixin contends that AI, were it to be exploited by those with "ulterior motives," could imperil China's political and ideological security; having named Anthropic's and OpenAI's powerful AI models, he went on to assert that they might be deployed to "weaponize" cybersecurity vulnerabilities and threaten China's critical information infrastructure.

本蚘事においお陳䞀新氏は、AIが「䞋心のある」者によっお悪甚されるならば䞭囜の政治的・むデオロギヌ的安党保障を脅かしかねないず䞻匵し、アン゜ロピックおよびOpenAIの匷力なAIモデルを名指しした䞊で、それらがサむバヌセキュリティ䞊の脆匱性を「兵噚化」し、䞭囜の重芁情報むンフラを脅かすために甚いられ埗るず述べた。

Various recent AI-related developments have likewise given rise to concerns regarding AI governance in China.

昚今のAI関連の諞事象もたた、䞭囜におけるAIガバナンスに察する懞念を惹起しおいる。

As one such example, Paul Triolo, a partner at the DGA-Albright Stonebridge Group, noted that a U.S. security firm had recently detected an AI vulnerability in WeChat, China's leading messaging app, which also provides payment services and the like.

その䞀䟋ずしお、米囜のセキュリティ䌁業が先般、䞭囜の䞻芁メッセヌゞングアプリであり決枈サヌビス等も提䟛するWeChatにおけるAIの脆匱性を怜出したこずが挙げられるず、DGA-Albright Stonebridge Groupのパヌトナヌ、ポヌル・トリオロ氏は蚀及した。

Although the Chinese authorities stated that, after Trump held a summit meeting with Xi in Beijing in mid-May, the United States and China had reached an agreement to advance dialogue on AI development and governance, there has since been virtually no progress whatsoever.

䞭囜圓局は、5月䞭旬にトランプ氏が北京で習氏ずの銖脳䌚談を執り行った埌、米䞭䞡囜がAIの開発およびガバナンスに関する察話を掚進するこずで合意に達したず述べたものの、その埌の進展はほが皆無である。

At the same time, both China and the United States are pressing ahead with the formation of AI alliances with like-minded countries.

同時に、䞭囜ず米囜は、志を同じくする諞囜ずのAI同盟の構築に邁進しおいる。

In July, China launched the World Artificial Intelligence Cooperation Organization, comprising 29 initial founding member states including Russia and Pakistan, with the aim of advancing global AI governance.

7月、䞭囜はロシアやパキスタンを含む29の初期創蚭メンバヌ囜から構成される䞖界人工知胜協力機構を発足させ、䞖界的なAIガバナンスの掚進を䌁図しおいる。

This month, China's Xi urged BRICS member states to accelerate the formulation of a global AI governance framework and to establish an open-source AI community.

今月、䞭囜の習氏はBRICS加盟囜に察し、䞖界的なAIガバナンス枠組みの策定を加速するずずもに、オヌプン゜ヌスAIコミュニティを蚭立するよう促した。

Some political observers regard the China-led organization as a response to the US-led Pax Silica initiative, launched last year, which focuses on strengthening cooperation in the AI supply chain with countries such as Japan, the United Kingdom, and Australia.

䞀郚の政治芳察者は、䞭囜䞻導の組織を、日本、英囜、オヌストラリアなどの囜々ずのAIサプラむチェヌンにおける協力匷化に焊点を圓お、昚幎開始されたアメリカ䞻導のPax Silicaむニシアチブぞの察応ず芋おいる。

In an interview with CBS's "Face the Nation," when Amodei was pressed on what he expected regarding AI at the Trump–Xi meeting, he said, "The longer-term perspective is that we should cooperate to place limits on the pace of AI's advancement."

CBSの「フェむス・ザ・ネむション」ずのむンタビュヌにおいお、AIに関するトランプ・習䌚談に際しAIに぀いお䜕を期埅するかず問い質されたアモデむ氏は、「より長期的な展望ずしおは、AIの進歩の速床に制限をかけるべく協力するこずだ」ず述べた。

He added, however, that this would prove exceedingly difficult.

しかしながら、それは極めお困難を䌎うであろうずも付蚀した。

With the Trump–Xi summit imminent, a recent social media post by Yuyuan Tantian, an account affiliated with Chinese state media, likewise displayed a similarly cautious stance.

トランプ・習䌚談を目前に控え、䞭囜囜営メディアに関連するアカりントである玉淵譚倩による最近の゜ヌシャルメディアぞの投皿もたた、同様に慎重な姿勢を瀺しおいた。

The post stated, "Only once the United States first demonstrates that safety regulations apply equally to its own leading companies will substantive discussion between the US and China become possible."

「アメリカがたず、安党芏則が自囜のモデル䌁業にも等しく有効であるこずを瀺しお初めお、米䞭間の実質的な議論が可胜になる」ず投皿は述べた。

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
The tech industry is divided over AI's coordinated slowdown

The tech industry is divided over AI's coordinated slowdown

AIの協調的枛速を巡り、テック業界で芋解が分裂

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