Loading your language..
The world's AI safety strategy depends on cooperation between the US and China, but the two sides view each other as a problem.

The world's AI safety strategy depends on cooperation between the US and China, but the two sides view each other as a problem.

C1🇯🇵 日本語 → 🇺🇞 English

September 17th, 2026

The world's AI safety strategy depends on cooperation between the US and China, but the two sides view each other as a problem.

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

As concerns grow over the risks posed by artificial intelligence, hopes for any kind of global initiative hinge on cooperation between the United States and China.

人工知胜がもたらすリスクぞの懞念が高たる䞭、䜕らかのグロヌバルな取り組みぞの期埅は米䞭の協力にかかっおいる。

However, these two superpowers appear to have grown skeptical of each other's AI strategies.

しかし、この超倧囜同士は盞手のAI戊略に察し懐疑的になっおいるように芋える。

AI governance is expected to be on the agenda at next week's scheduled meeting 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 gaining urgency in both countries, and American technology leaders say that some cooperation with China is essential to averting the worst-case scenario of uncontrollable AI.

この問題は䞡囜で緊急性を増しおおり、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だず述べおいる。

However, the two countries are competing for decisive superiority in AI, and Trump has warned that any attempt to regulate this technology would benefit China.

しかし、䞡囜はAIの決定的な優䜍性を争っおおり、トランプ氏はこの技術を芏制するいかなる詊みも䞭囜を利するず譊告しおいる。

As a result, any agreement reached at the talks may prove superficial, yet it could serve as a starting point for something more substantial.

そのため、䌚談での合意は衚面的なものになるかもしれないが、より倧きなものぞの出発点になり埗る。

"It may be nearly impossible to reach a formal agreement or consensus, 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に達するこず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。

"Trump and Xi simply need to create political space by acknowledging that AI poses risks to both countries."

「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、政治的な空間を生み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 rules could enable China 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、そうした芏制は䞭囜がAIの分野でアメリカを远い抜く芁因になり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 stated on social media, "We are leading China and every other country."

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䞻導しおいる」ず述べた。

The two countries are advancing this technology through distinctly different approaches.

䞡囜はそれぞれ異なる手法でこの技術を進展させおいる。

American companies are pushing the frontiers of AI, leveraging computing power 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歊噚に、AIのフロンティアを远求しおいる。

China, by contrast, aims to spread more affordable models worldwide through open source.

䞀方、䞭囜はオヌプン゜ヌスを通じお、より手頃な䟡栌のモデルを䞖界的に普及させるこずを目指しおいる。

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

䞭囜のAIモデルは目芚た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぀぀ある。

While most American frontier models are closed-source, with their design elements kept undisclosed, US Treasury Secretary Scott Bessent said on Tuesday that more open-source models need to be developed in order to compete with China.

アメリカのフロンティアモデルの倧半はクロヌズド゜ヌスで蚭蚈芁玠が非公開ずされおいるが、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 accuses China of attempting 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」匕き出そうずしおいるず非難しおいる。

In response, China has hit back, accusing the United States of trying to suppress Chinese companies in order to establish an "AI industry monopoly."

これに察し䞭囜は反撃し、アメリカが「AI産業の独占」を築くために䞭囜䌁業を抑圧しようずしおいるず非難しおいる。

David Sacks, Trump's senior adviser on AI, stated that China has "already refused" to slow down the pace of developing cutting-edge forms of AI.

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

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

同氏はフォックスニュヌスの「むングラハム・アングル」で、冷戊時代の゜連ずの栞緊匵ずは異なり、䞭囜がAIの玄束を守ったかどうかを怜蚌するこずは䞍可胜だず述べた。

Sachs said, "In this case, there is no trust and no verification," and continued, "So let's not be foolish and hand over the entire game to China."

サックス氏は「このケヌスでは信頌がなく、怜蚌もできない」ず述べ、「だから愚かになっお、詊合党䜓を䞭囜に匕き枡すのはやめよう」ず続けた。

Nevertheless, people with ties to the Chinese government regard cooperation on AI as beneficial.

それにもかかわらず、䞭囜政府ず関係を有する人々は、AIに関する協調が有益であるず芋なしおいる。

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

北京で開催された幎次安党保障・防衛䌚議にお、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」分野であり、北京ずワシントンの察話を進めるべきだず述べた。

"Because there is a need for cooperation, an ever-growing need," he remarked at the Xiangshan Forum.

「なぜなら協力ぞの必芁性、増倧する必芁性があるからだ」ず同氏は銙山フォヌラムで述べた。

"So how can we manage the balance?"

「では、どのようにバランスを管理できるのか」

Two major American AI companies—Anthropic and OpenAI—have each restricted access in China, asserting that the United States has a national security responsibility to maintain its lead in AI technology.

アメリカの䞻芁AI䌁業2瀟——アン゜ロピックずOpenAI——はそれぞれ䞭囜でのアクセスを制限しおおり、アメリカにはAI技術のリヌドを維持する囜家安党保障䞊の責務があるず䞻匵しおいる。

However, they are also calling for cooperation with China in order to manage safety.

しかし、安党性を管理するための䞭囜ずの協力も呌びかけおいる。

Last weekend, in an essay calling for a slowdown in AI development on safety grounds, Anthropic CEO Dario Amodei stated that the United States should slow China's progress, which includes maintaining American restrictions on China's purchases of the world's most advanced AI chips.

先週末、安党䞊の理由からAI開発の枛速を求める゚ッセむの䞭で、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らせるべきであり、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, "Cooperation with China, an authoritarian state that possesses far more advanced AI capabilities, is indispensable for a global adjustment of pace."

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である䞭囜ずの協力が䞍可欠だ」ず付け加えた。

His essay quickly drew criticism from Beijing, and on Monday the Chinese Foreign Ministry stated, "Stoking fear, inciting confrontation, and engaging in malicious competition benefit no one and only serve to disrupt the global AI governance process."

同氏の゚ッセむはすぐに北京から批刀を济び、䞭囜倖務省は月曜日、「恐怖を煜り、察立をあおり、悪質な競争をするこずは、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だけだ」ず述べた。

China's state-run Global Times went even further, calling his remarks "packed with containment clauses targeting China, essentially a 'Cold War playbook' for the AI field."

䞭囜囜営の環球時報はさらに螏み蟌み、同氏の発蚀を「䞭囜を暙的にした封じ蟌め条項が満茉で、本質的にはAI分野の『冷戊の手匕曞』だ」ず呌んだ。

Trump has been in discussions with Jensen Huang, the CEO of NVIDIA, and even phoned the world's leading computer chipmaker during a podcast interview on Monday.

トランプ氏はNVIDIAのCEOであるゞェンスン・フアン氏ず協議しおおり、月曜日のポッドキャストむンタビュヌ䞭に、䞖界最先端のコンピュヌタチップメヌカヌに電話をかけたこずさえある。

NVIDIA has generally supported chip exports to China.

NVIDIAは抂しお、䞭囜ぞのチップ茞出を支持しおいる。

This is seen as a way to build a more open relationship regarding AI, given that it shapes an American ecosystem for the technology and that China is home to many researchers.

それは、この技術のアメリカの゚コシステムを圢成し、䞭囜に倚くの研究者がいるこずを考慮すれば、AIに関しおより開かれた関係を築く方法ずしおである。

In both the United States and China, the increasingly powerful capabilities of AI are putting senior executives and policymakers on their guard.

米䞭䞡囜においお、AIのたすたす匷力化する胜力は、最高幹郚や政策立案者たちを譊戒させおいる。

In an article published on Sunday, the head of China's intelligence agency, the Ministry of State Security, warned that AI poses numerous threats to the nation, ranging from political and ideological security to cyberattacks.

䞭囜の諜報機関である囜家安党省のトップは日曜日の蚘事で、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、囜に倚くの脅嚁をもたらすず譊告した。

In this article, Chen Yixin argues that AI, if used by those with "ulterior motives," could threaten China's political and ideological security, and he named the powerful AI models of Anthropic and OpenAI.

本蚘事で陳䞀新氏は、AIが「䞋心のある」者によっお䜿甚されれば䞭囜の政治的・むデオロギヌ的安党保障を脅かす恐れがあるず䞻匵し、アン゜ロピックずOpenAIの匷力なAIモデルを名指しした。

According to him, they could be used to "weaponize" cybersecurity vulnerabilities and threaten China's critical information infrastructure.

同氏によれば、それらは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かねないずいう。

Concerns surrounding AI governance in China have been further heightened by several other recent AI-related developments.

䞭囜におけるAIガバナンスをめぐる懞念は、最近の他のいく぀かのAI関連の出来事によっお䞀局高たっおいる。

According to Paul Triolo, a partner at DGA-Albright Stonebridge Group, these include the fact that a U.S. security firm recently discovered an AI-related vulnerability in WeChat, a major Chinese messaging app that also offers services such as payments.

DGA-Albright Stonebridge Groupのパヌトナヌであるポヌル・トリオロ氏によれば、それには、米囜のセキュリティ䌁業が先ごろ、䞭囜の䞻芁メッセヌゞングアプリであり決枈サヌビスなども提䟛するWeChatにAI䞊の脆匱性を発芋したこずが含たれるずいう。

Chinese authorities stated that after Trump held a summit meeting with Xi in Beijing in mid-May, the United States and China agreed to advance dialogue on AI development and governance, but little progress has been seen since then.

䞭囜圓局は、5月䞭旬にトランプ氏が北京で習氏ず銖脳䌚談を行った埌、米䞭䞡囜がAI開発ずガバナンスに関する察話を進めるこずで合意したず述べたが、その埌ほずんど進展は芋られおいない。

At the same time, China and the United States are working to build AI alliances in cooperation with like-minded countries.

同時に、䞭囜ずアメリカは、志を同じくする囜々ず連携したAI同盟の圢成に取り組んでいる。

In July, China launched the World Artificial Intelligence Cooperation Organization, comprising 29 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なる䞖界人工知胜協力機構を立ち䞊げ、䞖界的なAIガバナンスを掚進するこずを目的ずしおいる。

This month, China's Xi called on BRICS member countries to accelerate the establishment of a global AI governance framework and to set up 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蚭定を加速し、オヌプン゜ヌスAIコミュニティを蚭立するよう呌びかけた。

Some political observers regard the China-led organization as a response to the US-led Pax Silica initiative, which was launched last year.

䞀郚の政治芳察者は、䞭囜䞻導の組織を、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。

This initiative focuses on strengthening cooperation in the AI supply chain with countries such as Japan, the United Kingdom, and Australia.

このむニシアチブは、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たものである。

In an interview with CBS's "Face the Nation," when asked what he expected from a Trump-Xi meeting on AI, Amodei said, "The longer-term thing is to cooperate to put limits on the rate of AI progress."

CBSの「フェむス・ザ・ネむション」ずのむンタビュヌで、AIに関するトランプ・習䌚談に䜕を期埅するかず問われ、アモデむ氏は「より長期的なこずは、AIの進歩の速床に制限をかけるために協力するこずだ」ず述べた。

However, he also remarked that this would be extremely difficult.

しかし、それは非垞に難しいだろうずも述べた。

Ahead of the Trump-Xi meeting, a recent social media post by Yuyuan Tantian, an account affiliated with Chinese state media, was similarly cautious in tone.

トランプ・習䌚談を控え、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が最近゜ヌシャルメディアに投皿した内容も、同様に慎重なものだった。

The post stated, "Only once the United States first demonstrates that safety rules 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查尔斯囜王圚安党担忧䞭䌚见人工智胜领袖

C1Sep 18
UN Secretary-General, in a turbulent world, speaks of the path forward before stepping down

UN Secretary-General, in a turbulent world, speaks of the path forward before stepping down

囜連事務総長、揺れる䞖界で退任前に前進の道を語る

C1Sep 18
Democratic frontrunners race to address AI threats amid Trump's denials

Democratic frontrunners race to address AI threats amid Trump's denials

民䞻党有力候选人圚特朗普的吊定声䞭竞盞应对AI嚁胁

C1Sep 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

C1Sep 18
OpenAI reports concerning new AI behavior, promises stricter tracking

OpenAI reports concerning new AI behavior, promises stricter tracking

OpenAI reports concerning new AI behavior, promises stricter tracking

C1Sep 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

C1Sep 17
Tech industry divided over calls for coordinated AI slowdown

Tech industry divided over calls for coordinated AI slowdown

Tech industry divided over calls for coordinated AI slowdown

C1Sep 17
President Trump downplayed the need for oversight of AI development, stating he does not want to cede superiority to China.

President Trump downplayed the need for oversight of AI development, stating he does not want to cede superiority to China.

トランプ倧統領、AI開発の監芖の必芁性を軜芖し、䞭囜に優䜍を譲りたくないず発蚀

C1Sep 14
Oprah Winfrey hopes you'll experience an "AHA" moment at Sphere

Oprah Winfrey hopes you'll experience an "AHA" moment at Sphere

オプラ・りィンフリヌ、スフィアで「AHA」の瞬間を䜓隓しおほしいず願う

C1Sep 14