Alibaba falls further amid accusations tied to Anthropic AI model security concerns

Big tech fight over AI secrets

A major clash is unfolding between two tech giants. Anthropic, a U.S. AI company, has accused Alibaba-affiliated operators of using fraudulent accounts to improperly access Claude, Anthropic’s AI model family.

Anthropic says the activity was a large-scale distillation campaign designed to extract Claude’s capabilities and improve rival AI systems. Alibaba did not immediately respond to Reuters’ request for comment.

What Alibaba is accused of?

The accusation centers on something called distillation attacks. In simple terms, this means using a powerful AI model to train a weaker one. Anthropic says Alibaba operators targeted their most advanced AI skills.

The company claims the campaign was massive. Between April and June, there were 28.8 million exchanges with Claude. That’s Anthropic’s flagship AI system. Nearly 25,000 fake accounts were reportedly used to pull off this plan.

Huge scale of alleged attack

The numbers involved are mind-boggling. Think about 28.8 million interactions happening in just weeks. That’s more than many websites see in a year. Anthropic says this is the largest known attack of its kind against them.

The fake accounts made this possible. They let Alibaba operators bypass normal restrictions. Anthropic believes this was a coordinated effort. It wasn’t just a random activity but a systematic attempt to harvest their technology.

Why does this matter to you?

You might wonder why this matters if you aren’t in tech. Anthropic argues that illicit distillation can create unfair advantages because competitors may gain advanced capabilities faster and at lower cost than building them independently.

The company also warns that illicitly distilled models may not keep the safeguards built into the original systems. That could make copied models riskier if they are used for cyber, military, surveillance, or other sensitive purposes.

Fun fact: Analysts warn this could become a new AI supply-chain risk, where copied AI models lose the safety controls they were built with.

Anthropic’s letter to lawmakers

Anthropic didn’t stay quiet about this. They wrote a letter to two US senators. The letter was sent to Senator Tim Scott and Senator Elizabeth Warren. They’re the leaders of the Senate Banking Committee. The company wanted Congress to know what happened.

In the letter, Anthropic asked for help. They want new laws to stop these attacks. They also asked for penalties against companies that do this. The company believes the government needs to step in to protect American innovation.

Fun fact: The letter came less than two months after the White House issued a memo pledging to help AI companies fight mass distillation attacks.

White house gets involved

The White House had already moved on the broader issue before Anthropic’s letter became public. In an April 23 memo, the Office of Science and Technology Policy said foreign entities, principally based in China, were engaged in industrial-scale campaigns to distill U.S. frontier AI systems.

The memo said the Trump administration would share information with U.S. AI companies, help the private sector coordinate, develop best practices, and explore measures to hold foreign actors accountable. Anthropic’s accusations against Alibaba landed inside that wider U.S. push to protect American AI technology.

How does AI Distillation work?

Let’s break down what distillation means. Imagine you have a brilliant professor. You want to teach a student everything the professor knows. Instead of doing the hard work yourself, you record all the professor’s lectures. Then you have the student study those recordings.

In AI, it works similarly. A less advanced model learns from a more advanced one. This can be legitimate if done with permission. But Anthropic says Alibaba did this without asking. It’s like taking someone’s work and claiming it as your own.

Alibaba’s response so far

Alibaba has not publicly issued a detailed response to Anthropic’s allegations. Reuters and Business Insider reported that Alibaba did not immediately respond to requests for comment.

The company is also challenging a separate Pentagon designation that placed it on a list of Chinese military companies. Alibaba says the designation has no basis in fact or law and argues that its products and services are built for retail, logistics, and enterprise technology, not defense or intelligence.

Alibaba stock takes a hit

The news hit Alibaba’s stock hard. Shares dropped about 5% in Hong Kong trading. They hit their lowest point since early 2025. This is tough news for investors who believed in the company. The stock has already lost about 33% of its value this year.

Other Chinese tech companies also felt the impact. Baidu and Xiaomi both saw their stocks fall, too. Investors are nervous about the future of Chinese AI companies. They worry that US rules might block them from key technologies.

Big impact on US-china relations

This dispute comes at a tense time in the U.S.-China AI race. The White House says the United States leads in AI, but it has also warned that foreign entities, principally based in China, are using coordinated distillation campaigns to extract capabilities from American frontier models.

Anthropic argues that these campaigns can let foreign competitors gain powerful capabilities faster and at lower cost than building them independently. That is why the issue has moved beyond a private company dispute and into U.S. national security policy.

The irony in this complaint

There is a complicated twist here. The U.S. government issued a June 12 export-control directive that forced Anthropic to suspend access to Fable 5 and Mythos 5 for all users because it could not verify users’ nationalities in real time.

Those restrictions were later lifted by the Commerce Department after Anthropic added safeguards, though Mythos 5 access remained limited to selected approved U.S.-based organizations. That means Anthropic is asking for protection against unauthorized model extraction while also navigating federal scrutiny of its own most advanced systems.

If you want to see how China’s AI ambitions are still moving forward despite growing geopolitical pressure, check out how a Chinese humanoid robot startup gets a $100M boost led by Alibaba for a closer look at what’s powering the next wave of competition.

What comes next in this saga?

The next step is likely to play out in Washington as well as in the market. A House bill called the Deterring American AI Model Theft Act of 2026 would create a process for identifying foreign model-extraction attackers and making them eligible for export blacklist action or sanctions.

Alibaba is also fighting the Pentagon’s Chinese military company designation in federal court while facing Anthropic’s separate AI allegations. The case shows how quickly the AI race is becoming a fight over access, national security, and control of frontier technology.

If you want to see how Alibaba is still pushing ahead with new AI tools despite rising political pressure, check out Alibaba unveils AI agent to accelerate online sourcing for merchants for a closer look at its latest move in the global AI race.

If you want to see how Alibaba is still pushing ahead with new AI tools despite rising political pressure, check out Alibaba unveils an AI agent to accelerate online sourcing for merchants for a closer look at its latest move in the global AI race.

What’s your take on this AI showdown? Sound off in the comments, and don’t forget to like this post if you want more updates like this.

This slideshow was made with AI assistance and human editing.

Don’t forget to follow us for more exclusive content on MSN.

Read More From This Brand:

• Google lands $200 billion cloud deal with Anthropic
• Anthropic unveils civilian Mythos amid safety debate
• Anthropic may ditch Nvidia for Microsoft’s custom AI hardware