Anthropic, the U.S. artificial intelligence company and OpenAI’s closest rival, has accused China’s Alibaba of orchestrating a large-scale attempt to steal its proprietary AI model capabilities, escalating tensions between Washington and Beijing over technology espionage. In a series of statements issued on Thursday, Anthropic alleged that Alibaba used fake accounts to illicitly extract the functionalities of its Claude AI system, a claim the Chinese e-commerce giant has not publicly addressed.

The accusations, first reported by the Financial Times and corroborated by multiple international outlets, mark one of the most direct allegations of industrial espionage in the global AI sector. Anthropic’s chief executive, Dario Amodei, stated that the company had detected coordinated efforts to reverse-engineer its models through repeated queries from fabricated user profiles. “This was not a random probe—it was a systematic campaign,” Amodei told reporters. The company has since implemented additional safeguards and notified U.S. authorities, including the Department of Commerce, which is investigating the incident under export control regulations.

The timing of the allegations is politically sensitive. Alibaba filed a lawsuit against the Pentagon on the same day, challenging its inclusion on a blacklist of Chinese companies alleged to have ties to the military . The company argues the designation violates constitutional due process, framing the dispute as part of a broader U.S. strategy to restrict Chinese technological advancement. Meanwhile, Anthropic’s claims risk deepening the rift between the two superpowers, already strained by export controls on advanced semiconductors and restrictions on foreign investment in AI firms.

Industry analysts warn that such incidents could accelerate a bifurcation of the global AI ecosystem. “We’re seeing the emergence of two distinct AI supply chains—one centered in the U.S. and its allies, the other in China,” said Kai-Fu Lee, CEO of Sinovation Ventures. “This is no longer just about competition; it’s about control.” The European Union, which has sought to position itself as a neutral hub for AI development, has yet to comment on the dispute.

Anthropic’s allegations come amid broader concerns over AI security. Earlier this week, Palo Alto Networks revealed it had been among the first companies to test Mythos, Anthropic’s most advanced model, and described its capabilities as “both revolutionary and alarming” . The company’s experience underscores the dual-use nature of cutting-edge AI systems, which can be exploited for both innovation and misuse.

As the investigation unfolds, the incident raises urgent questions about the adequacy of current safeguards against AI model theft. With geopolitical stakes higher than ever, the outcome could redefine the rules of engagement in the global AI race.

Bluesky Threads

Source Intelligence

10 sources5 countries

Geographic Origin8 located

2 further sources not geolocated

Political Spectrum7 mapped

Articles

KI-Unternehmen Anthropic wirft Alibaba-Konzern Kopier-Attacke vor Der chinesische Konzern Alibaba soll versucht haben, widerrechtlich Fähigkeiten von Anthropic-Software mit Künstlicher Intelligenz abzugreifen.

die presse · about 2 hours ago

GOKO Showcases M6 Robotic Lawn Mower at spoga+gafa 2026, Built for Slopes, Rough Ground, and Large Yards KENT, ENGLAND - Media OutReach Newswire - 25 June 2026 - At spoga+gafa 2026, GOKO is debuting the M6 AI-powered robotic lawn mower, built for large and challenging yards. With 4WD mobility, AI quad-vision cameras, RTK + VSLAM navigation, and powerful cutting performance, the M6 handles slopes, ...

the slovenia times · about 2 hours ago

KI-Konzern: Anthropic wirft chinesischem Konzern Kopier-Attacke vor Die KI-Firma Anthropic gilt als schärfster Rivale des ChatGPT-Entwicklers OpenAI. Jetzt könnten ihre Vorwürfe in Richtung China für neuen Zündstoff zwischen Washington und Peking sorgen.

handelsblatt · about 2 hours ago

Alibaba sues Pentagon over blacklist designation Chinese e-commerce giant Alibaba has filed a lawsuit against the US Department of Defense over its inclusion on a blacklist identifying companies with alleged ties to the Chinese military. The company argues the designation violates constitutional due process and free speech rights. The lawsuit follows the Pentagons recent addition of several Chinese tech firms to the list.

yenisafak · about 2 hours ago

YouTube settles teen lawsuit over addictive design: Report Googles YouTube has settled a lawsuit brought by a Florida teenager who alleged the platforms addictive design features caused mental health harm, as the case becomes part of broader litigation involving over 1,000 similar claims against social media giants scheduled for trial in July.

yenisafak · about 2 hours ago

Aberdeen Technical School "ATS Robot" Wins 1st Runner-Up at Robofest World Championship Peak Showdown Against National Team Spotlights Innovation and Technology Friendship HONG KONG SAR - Media OutReach Newswire - 25 June 2026 - The robotics team from Aberdeen Technical School (ATS), ATS Robot, has shone brightly on the academic innovation and technology stage in recent years. The ...

the slovenia times · about 3 hours ago

Striding AI Announces Plans to Develop Next-Generation Robotic Foundation Systems for Physical AI Deployment BEIJING, CHINA - Media OutReach Newswire - 25 June 2026 - Striding AI today announced that it is developing a new generation of robotic foundation systems designed to accelerate the deployment of Physical AI in real-world environments. Striding AI Announces Plans to Develop Next-Generation ...

the slovenia times · about 3 hours ago

AI-powered legal win sets landmark precedent in UK courts In a historic first for British jurisprudence, an artificial intelligence system successfully prepared a legal case that resulted in a favourable ruling. The claimant, who was represented in court by a human barrister, recovered £7,000 in unpaid dues, signalling a transformative moment for affordable legal services.

yenisafak · about 3 hours ago

Phase 4 of the International Volunteer Project: Dive Against Debris®, Beach Cleanup & Environment Education At Pulau Redang, Terengganu Darul Iman (27-29 June 2026) PULAU REDANG, MALAYSIA - Media OutReach Newswire - 25 June 2026 - The Black Eye Scuba Team ("BEST"), in collaboration with the Jabatan Pendidikan Negeri Terengganu, Agensi Penguatkuasaan Maritim Malaysia, Jabatan BOMBA & Penyelamat Negeri Terengganu, Angkatan Pertahanan Awam Malaysia (APM) ...

the slovenia times · about 3 hours ago

Anthropic accuses Alibaba of large-scale AI model extraction

news.yahoo.com · about 3 hours ago

Impossible Marketing Unveils ImpossiblePlus™ AI SEO Solution for Singapore Businesses SINGAPORE - Media OutReach Newswire - 25 June 2026 - Impossible Marketing, Singapores leading SEO agency and pioneer of AI SEO, has announced the launch of ImpossiblePlus™ - a proprietary AI visibility platform built to help businesses track how their brand is being recommended, cited, and ...

the slovenia times · about 3 hours ago

„Război total. În culisele luptei pentru putere între cele mai puternice două companii din AI Dacă nu ar fi existat rivalitatea intensă dintre Anthropic și OpenAI, este posibil ca boom-ul inteligenței artificiale generative să nu fi venit atât de repede. La sfârșitul anului 2022, OpenAI a aflat că Anthropic lucra la un chatbot bazat pe inteligență artificială. Reuters relatează că Sam Altman, directorul OpenAI, le-a cerut imediat angajaților să accelereze …

hotnews.ro · about 3 hours ago

Anthropic wirft Alibaba KI-Spionage vor

orf.at · about 3 hours ago

„Un moment «tanc versus cal» al secolului nostru. AI reduce planificarea operațiunilor în armata britanică de la 72 de ore la o oră Armata britanică a reușit să reducă ciclurile de planificare a operațiunilor la nivel de corp de armată de la 72 de ore la doar o oră, datorită unui nou sistem bazat pe inteligență artificială, a declarat șeful Statului Major General, generalul Roly Walker, relatează Business Insider.

digi24 · about 3 hours ago

AI-topparnas varning: Sverige tappar jobben Sveriges hetaste AI-bolag har fått nog. Grundarna bakom Lovable, Legora och Tandem Health går till regeringen med hård kritik mot svenska regler – och får med sig Marcus Wallenberg. Nu varnar de: framtidens jobb kan hamna utomlands.

svenska dagbladet · about 3 hours ago

Künstliche Intelligenz: Anthropic wirft chinesischem Konzern Kopier-Attacke vor

die zeit · about 5 hours ago