Connect with us

America

Anthropic accuses Chinese AI labs of ‘industrial-scale’ theft to bypass US export curbs

Published

on

Anthropic has leveled blistering accusations against three prominent Chinese artificial intelligence laboratories, alleging “industrial-scale” campaigns of intellectual property theft that the startup claims undermine US national security and circumvent rigorous export controls.

The San Francisco-based creator of the Claude AI interface asserted Monday that DeepSeek, Moonshot, and MiniMax have engaged in “industrial-scale distillation attacks” against its proprietary models.

Distillation refers to the practice of training smaller, leaner AI systems on the outputs of more sophisticated models. This technique allows developers to replicate elite performance levels without requiring the massive computational infrastructure typically necessary to achieve such benchmarks.

The issue has gained significant geopolitical resonance as Chinese AI groups grapple with tightening US export restrictions. These controls limit access to Nvidia’s most advanced hardware, including the high-end Blackwell series, forcing firms in the world’s second-largest economy to pivot toward alternative strategies. To mitigate the impact of the silicon blockade, Chinese entities have reportedly turned to offshore model training, the utilization of legacy or black-market semiconductors, and aggressive engineering efficiencies to drive down costs.

Anthropic disclosed it has detected 24,000 fraudulent accounts linked to the effort, documenting more than 16 million interactions with Claude. The company alleges these communications were systematically engineered to “train and refine” the Chinese firms’ internal models.

“Distillation attacks allow foreign laboratories, including those under the control of the Chinese Communist Party, to erode the competitive advantage that export controls were designed to protect, effectively neutralizing those safeguards through other means,” Anthropic stated.

Over the past month, Chinese AI groups including Moonshot and MiniMax have debuted a suite of high-performance models. These releases have earned praise from developers for their efficiency in powering AI agents and video generation applications.

DeepSeek, which stunned Silicon Valley in January with its breakthrough R1 model—celebrated for matching Western performance at a fraction of the cost—had been widely expected to unveil a new iteration ahead of last week’s Lunar New Year holiday. However, the company has yet to issue a formal announcement.

Since the debut of R1, DeepSeek has focused on incremental updates rather than a new flagship release. This strategic pause has allowed domestic rivals such as Alibaba and ByteDance to gain ground, attracting developers seeking low-cost, open-source alternatives.

US artificial intelligence firms and Washington officials have intensified warnings that distilled models present acute national security risks.

Anthropic argued that models “created through illicit distillation” are unlikely to retain the rigorous safety guardrails designed to prevent AI from being weaponized for biological warfare or malicious cyber operations.

The allegations follow a similar move by OpenAI in January. The ChatGPT creator revealed it had uncovered evidence that its foundational models were being distilled, pointing to DeepSeek as the suspected beneficiary.

Earlier this month, OpenAI submitted a memorandum to the US House Select Committee on Strategic Competition between the United States and the Chinese Communist Party. The memo argued that DeepSeek’s upcoming model must be viewed within the context of “efforts to leverage capabilities developed by OpenAI and other US frontier laboratories for free.”

The startup further accused DeepSeek of reflecting “CCP censorship and information control,” maintaining that “adversarial distillation poses serious cost, security, commercial, and strategic risks for the US.”

America

Trump administration targets 60 nations with new tariff draft under Section 301

Published

on

The US administration is proposing new tariffs of at least 10% on imports from 60 trading partners, following an investigation into goods allegedly produced using forced labor.

According to a Bloomberg report citing sources within the Office of the US Trade Representative (USTR), the specific tariff rates will vary based on individual countries’ legislative frameworks regarding forced labor and their capacity to enforce those laws.

Under the drafted regulations, a 10% tariff rate will apply to imports from the European Union, Mexico, Canada, the United Kingdom, Taiwan, and several other nations. Conversely, goods arriving from China, India, Japan, South Korea, Switzerland, and Brazil will be subject to a 12,5% tariff.

The USTR stated that the lower tariff rate will apply to products from nations that prohibit forced labor or have committed to doing so. The agency emphasized that states failing to establish such prohibitions or lacking the capacity to effectively enforce them will face the higher tariff rate.

Bloomberg reported that this step represents a continuation of President Donald Trump’s policy to reinstate across-the-board tariffs on all countries, which had previously been ruled unconstitutional.

The proposed tariffs are the result of investigations initiated under Section 301 of the Trade Act of 1974.

Commenting on the development, Deborah Elms, Head of the Trade Policy Group at the Hinrich Foundation in Singapore, said, “This is highly significant because Section 301 is an extremely powerful tool and is highly unlikely to be overturned. This opens the door to a range of new tariff and non-tariff measures.”

The report noted that the tariffs are being introduced at what could be a turning point for the global economy.

Financial markets are already navigating a sensitive period due to rising gas and oil prices driven by conflict in Iran.

The new tariffs will not take effect immediately. Before implementation, a review and evaluation period will be conducted, which may lead to modifications in the draft proposal.

According to the timeline reported by Bloomberg, written comments on the tariffs must be submitted by July 6. Additionally, the Section 301 Committee is scheduled to hold a public hearing on July 7.

US Trade Representative Jamieson Greer argued that forced labor practices in partner nations force American workers to compete on an unequal playing field. “We will no longer tolerate this unfairness,” Greer said.

On the other hand, the USTR proposed certain tariff exemptions that could affect apparel and textile imports. While these goods could enter the US at reduced tariff rates, quotas would be determined based on the respective countries’ existing textile exports to the US.

Beef, tomatoes, bananas, coffee, orange juice, and several other food products will be entirely exempt from the tariffs. Furthermore, double taxation will not be imposed on metals, specific fuel types, and chemicals that are already subject to other duties.

In May, the US Court of International Trade ruled that the 10% tariff on foreign imports promoted by President Donald Trump was unlawful. Defending the White House’s objectives following the court ruling, Trump characterized the judges as “radical left-wing” and remarked, “Nothing surprises me. We always find different ways. We make a decision and act in another way.”

In February, the US Supreme Court also ruled that tariffs established by Trump were contrary to the law. The court concluded that the president had exceeded his authority in imposing those duties. Trump, however, claimed that the court was under foreign influence.

Continue Reading

America

Google seeks approval to release 32 million mosquitoes in US disease-control project

Published

on

Google is seeking federal approval to release nearly 32 million mosquitoes in California and Florida as part of a biological pest-control initiative known as the Debug project.

The little-known program aims to combat disease-carrying mosquitoes by releasing millions of sterile male mosquitoes into the environment, an approach designed to stop “bad bugs with good bugs.”

According to the US Centers for Disease Control and Prevention (CDC), mosquitoes are classified as the world’s deadliest animals. Of the more than 3,500 mosquito species that exist globally, only Aedes aegypti is responsible for transmitting dengue fever, Zika virus and chikungunya, diseases that sicken hundreds of millions of people each year.

In a statement published on the official website of the Debug project, Google described the issue as a difficult problem to solve, noting that many mosquito-borne diseases lack effective vaccines or treatments.

The statement argued that relying on pesticides is not a sustainable solution because such chemicals become less effective over time and can be toxic. It also said that eliminating standing water alone is insufficient because it is impossible to identify every breeding site used by mosquitoes.

For those reasons, Google said a new approach is required and that it found a solution in what it describes as “good” mosquitoes of the same species.

The project website explains the method as follows:

“Good bugs are the same mosquito species as the bad bugs that spread disease. Our good bugs are male mosquitoes carrying Wolbachia, a naturally occurring bacterium found in nature. This bacterium prevents them from producing offspring with wild female mosquitoes. Male mosquitoes do not bite and cannot spread disease, so the good bugs will stop the bad bugs from reproducing. Over time, fewer bad mosquitoes will remain.”

Scientists involved in the Debug project emphasized that the technique relies entirely on a naturally occurring bacterium, contains no chemicals or toxins, and does not involve genetic modification.

Researchers said similar approaches have been used safely for decades to control other pests. They added that the Debug team is combining scientific and engineering expertise with support from international partners in an effort to suppress disease-carrying mosquito populations.

Project scientists said their approach differs from previous eradication programs because it applies the Sterile Insect Technique on a larger scale through the use of data analytics, sensors and automation.

According to information published in the project’s frequently asked questions section, program officials are working closely with national and local governments, community leaders and research institutions.

Officials said they meet with residents in areas targeted for deployment before operations begin in order to better understand local concerns and priorities.

Google is therefore continuing to pursue federal authorization to implement the project in both California and Florida.

A notice published in the Federal Register shows that the US Environmental Protection Agency (EPA) is reviewing Google’s applications for an Experimental Use Permit under the Federal Insecticide, Fungicide, and Rodenticide Act.

According to details contained in the filing, nearly 16 million mosquitoes would be released in Florida during the first year of the project.

A further 16 million mosquitoes would be released in California during the second year.

Members of the public can obtain additional information and submit comments through the federal rulemaking portal by visiting regulations.gov and entering docket identification number EPA-HQ-OPP-2025-3951.

Continue Reading

America

US Marines test lower-cost counter-drone system to reduce missile dependence

Published

on

US Marine Corps personnel tested a new counter-drone defense system during military exercises held in the Philippines in April.

According to a report by The Wall Street Journal (WSJ), the system is designed to avoid the continuous use of expensive missiles and instead relies on a coordinated set of countermeasures.

The system consists of two armored vehicles known collectively as MADIS (Marine Air Defense Integrated System).

One vehicle is equipped with an advanced radar system, while the other carries the Stinger air defense missile system. Both vehicles are also fitted with a small cannon, a machine gun and electronic warfare equipment.

According to the report, MADIS is intended to provide military personnel with multiple options for engaging drones, including cannon fire, missiles and electronic warfare tools.

The objective is to reduce dependence on high-cost weapons when protecting military units and other strategic assets.

US Marine Corps officials told WSJ that one of the system’s most effective features is its ability to fire specially manufactured 30-millimeter ammunition equipped with precision fuzes that detonate as they approach a target.

Steven Sawyer, a former ammunition technician at the NATO Support and Procurement Agency, told the newspaper that 30-millimeter rounds are generally less accurate than missiles but are significantly cheaper to use.

Sawyer said that even if five such rounds were required to destroy a drone, the total cost would remain around $11,250.

By comparison, a single Stinger missile costs about $430,000, while Coyote interceptor missiles used in conflicts in the Middle East are priced between $100,000 and $125,000 each.

Sawyer added that 30-millimeter ammunition has proven effective against Shahed-family drones, which cannot be neutralized through electronic warfare methods.

At the same time, he stressed that US defense companies continue to face difficulties producing sufficient quantities of the ammunition. According to Sawyer, the precision fuzes are highly sophisticated electromechanical devices and only a limited number of manufacturers can produce them at scale.

WSJ noted that countering large numbers of inexpensive drones has become one of the most pressing challenges facing modern militaries.

The US military has encountered the problem directly during operations in the Middle East, where it has been forced to expend limited stocks of extremely costly precision-guided munitions.

Previously, the South China Morning Post (SCMP) reported that Chinese scientists had developed a combat algorithm known as HG-STR based on a “kill them all” concept.

The algorithm was said to enable swarms of fixed-wing drones to autonomously scan the battlefield and destroy enemy targets even if communications are disrupted and lines of sight are obstructed.

In April, The New York Times, citing three sources within defense and intelligence agencies, reported that the Pentagon assessed Russia’s and China’s drone development programs to be more advanced than those of the United States.

The assessment regarding China’s drone capabilities was reportedly based on analysis of a military parade held in China in September 2025.

Continue Reading

MOST READ

Turkey