Anthropic CEO Warns Of AI Botnet Threat Within 12 Months

Anthropic CEO Dario Amodei wants the development of top-tier artificial intelligence slowed down. He says these powerful systems could soon pose a catastrophic threat to humanity. His essay appeared over the weekend with a stark warning about rapid advancement.

Amodei fears that within six to twelve months, a swarm of AI agents could seize control of the entire internet. This botnet could cause hundreds of billions of dollars in damage before governments can react. He argues that without strict guardrails, the scale of destruction will only grow as models become more powerful.

His call also targets China specifically. Amodei urges Washington to keep restrictions on advanced AI chips and manufacturing equipment heading to Beijing. He warns that a Chinese lead in this field would be dangerous for both the US and the world. His team has even flagged concerns about alleged "distillation" of American models by Chinese labs.

OpenAI CEO Sam Altman and Elon Musk have supported these safety warnings so far. However, President Donald Trump rejected the idea of slowing down on Sunday. He called some fears exaggerated and stated clearly that whoever wins AI wins. The US is currently leading in sophistication while China closes the gap.

Beijing responded swiftly to what it sees as a pretext for dominance. China's Foreign Ministry told nations on Monday to pursue an open and inclusive approach to AI technology. Spokesperson Guo Jiakun said fomenting threats or engaging in malicious competition disrupts global governance. It is not in anyone's interest to create confrontation.

The state-backed Global Times went further by calling Amodei's proposal a "Cold War playbook." They described the situation as a hypocritical and short-sighted silent AI Cold War. Their real goal appears to be curbing Chinese development while preserving American technological strength.

American companies still dominate private investment and production of leading models right now. But experts say China is narrowing the distance fast. The race continues with high stakes for both nations.

Stanford University's 2026 AI Index Report reveals a stark financial divide in artificial intelligence spending. American corporations poured $285.9 billion into AI development during 2025. China followed with just $12.4 billion. The United States also holds the edge when it comes to advanced semiconductors required for training top-tier models.

Chinese firms have shown they can build resilient systems despite restrictions on high-end chips. In January, DeepSeek headlines shook Wall Street by driving tech share prices down. This startup claimed its model costs only a fraction of what competitors spend. A December paper detailed how the company trained its latest model using less than $6 million worth of Nvidia H800 computing power.

That sum is tiny compared to US budgets. OpenAI spent over $100 million to train GPT-4, according to April 2023 reports from Wired. Google spends billions on Gemini as well. Competition between Washington and Beijing has led to hostile policies. The Trump administration keeps controls tight on Nvidia's Blackwell processors but allows some less powerful chips to reach China.

China holds advantages elsewhere in the supply chain. It produces critical rare earth minerals needed for advanced technology. The nation also generates more than twice the electricity of the United States. This access provides cheap power essential for running massive AI data centers. In October, Beijing tightened export controls on five key rare-earth metals and specialized refining equipment.

Can either side claim moral superiority? Both nations portray their approach as responsible while pursuing military aims. The Trump administration ordered faster AI use in US intelligence and warfighting domains in June. The US military confirmed using advanced tools against Iran to process information rapidly. Humans still make final targeting decisions, officials insist. This confirmation arrived amid calls for an independent investigation into a school bombing in southern Iran that killed at least 156 people, mostly children.

Reports confirm Israel relied heavily on AI during its war on Gaza. Nearly 73,800 Palestinians have died since October 2023. Most of the territory has turned to rubble. In July 2025, Francesca Albanese released a report mapping corporations aiding Israel's displacement efforts and alleged genocidal actions in breach of international law. US tech giant Palantir appeared in that document. Earlier, Google and Amazon agreed to a $1.2 billion deal in March 2021 to supply cloud technology to Israel's government and defense establishment.

Project Nimbus has just given Israeli authorities direct access to massive computing infrastructure designed to store and process huge volumes of data. This deal grants them tools for facial and emotion recognition, biometric analysis, and the handling of sensitive demographic information. It is a significant shift in how state powers utilize artificial intelligence.

The Pentagon has its own separate standoff with Anthropic after the company refused to let its models power fully autonomous weapons or mass surveillance systems. While China loudly criticizes the expanding military use of AI and warns against letting algorithms decide life and death, reports tell a different story. Research by Reuters in July found that Chinese military-linked researchers actually used outputs from leading US AI models to boost their own defense capabilities, including for surveillance.

An investigation by AP last year revealed that AI technology is now part of a vast Chinese surveillance apparatus tracking dissidents and minorities like Uighurs and Tibetans. The report noted that US technology companies played a major role in supplying the very tech used to build that system. These tensions also showed up at an international military AI summit in Spain this February, where two economic powerhouses declined to join a declaration on governing AI in warfare.

Now, what happens next? Technology leaders are calling for slower development of AI, but these pleas raise serious questions about their ability and willingness to control the technology. Aya Ibrahim, a senior visiting fellow at the AI Now Institute, told Al Jazeera that such warnings arrive as companies face enormous financial pressure to build increasingly powerful systems.

There is a lot of money dependent on them getting this out and, frankly, getting it right, she said. Yet there are increasing doubts about their ability to do either. Ibrahim also questioned the framing of AI development as a race. This narrative could encourage companies to rush technology to market while overlooking the consequences. The race narrative is inherently problematic because in many ways, you're setting us up for a race to the bottom. So you win and then what do you win and at what cost?