标签： Asia

Tensions between the United States and China over artificial intelligence development flared up again in late April 2026, when the Trump administration formally pledged to curb what it labels industrial-scale, unauthorized intellectual property extraction from leading U.S. AI models, just one day before Chinese AI developer DeepSeek launched its latest high-performance frontier model, DeepSeek V4.

In an official memorandum issued by the White House Office of Science and Technology Policy (OSTP) on Thursday, April 23, OSTP Director Michael Kratsios detailed allegations that foreign entities, primarily based in China, have been running coordinated, large-scale campaigns to distill proprietary capabilities out of cutting-edge U.S.-developed AI systems. According to Kratsios, bad actors leverage networks of tens of thousands of fake proxy accounts to bypass platform detection, while using jailbreaking methods to access protected proprietary information hidden within these models. This systematic siphoning, Kratsios argued, lets foreign competitors exploit billions of dollars in U.S. research investment and innovation to build competing products.

While Kratsios noted that distilled models do not match the full performance of the original U.S.-built systems, he emphasized that they can achieve comparable results on common industry benchmarks at a tiny fraction of the development cost for their creators. Beyond commercial unfairness, Kratsios claimed these unauthorized distillation campaigns also let bad actors deliberately remove built-in safety protocols and alignment safeguards that ensure original AI models operate responsibly and remain ideologically neutral and fact-based.

To counter these alleged activities, the memorandum outlined a four-pronged strategy for the Trump administration: first, it will share timely intelligence with U.S. AI development companies detailing attempted large-scale unauthorized distillation, including specific tactics used and the actors behind them; second, it will facilitate tighter cross-sector coordination between private AI firms to collectively disrupt these campaigns; third, it will partner with industry to develop standardized best practices for detecting, mitigating, and remediating large-scale distillation attacks, while helping firms strengthen their defensive systems; and finally, the administration will explore new enforcement and policy measures to hold bad actors accountable for these activities.

The White House’s public warning landed just 24 hours ahead of DeepSeek’s April 24 launch of DeepSeek V4, a moment that highlights growing anxiety in Washington over the rapid pace at which Chinese AI developers are closing the performance gap with the United States’ top frontier models. Unlike many firms that keep training methods private, DeepSeek has been transparent about its use of knowledge distillation, a common training technique where a smaller “student” model learns from the outputs of larger, more capable “teacher” models. When the company launched its V3 model in January 2025, it openly confirmed it used knowledge distillation during training. For its new V4 model, DeepSeek says it has advanced the technique with a new method called On-Policy Distillation (OPD), which draws on outputs from 10 separate teacher models to refine its student model’s outputs, speeding up the training cycle significantly.

According to DeepSeek’s published research, its top-tier DeepSeek-V4-Pro-Max outperforms OpenAI’s GPT-5.2 and Google’s Gemini-3.0-Pro on standard industry reasoning benchmarks, while its lighter, more efficient DeepSeek-V4-Flash-Max matches the performance of those two leading U.S. models at a far lower development and inference cost. The company also noted that its V4 line’s overall performance is only 3 to 6 months behind the very latest frontier models from U.S. developers, OpenAI’s GPT-5.4 and Google’s Gemini-3.1-Pro.

The current friction around AI distillation is not new. The launch of DeepSeek V3 in 2025 sent ripples through global financial markets, as investors reacted to the emergence of a low-cost Chinese AI model that could compete head-to-head with leading U.S. offerings. During a Senate hearing that January, U.S. Commerce Secretary Howard Lutnick claimed DeepSeek had built its model “dirt cheap” by sourcing Nvidia AI chips through third-party countries and leveraging open training data from Meta. That said, then-recent comments from President Trump struck a different tone: in February 2025, he argued that the development of lower-cost AI was an inevitable technological shift that could ultimately benefit the United States, calling falling AI development costs “a very good development.”

Criticism of Chinese AI firms’ distillation practices faded for nearly a year before resurfacing in early 2026, when U.S. AI leaders published formal accusations of so-called “distillation attacks” that reignited the political debate. In a February 12 memo to the U.S. House Select Committee on China, OpenAI accused DeepSeek of using distillation techniques to “free-ride on the capabilities developed by OpenAI and other U.S. frontier labs,” adding that it had detected new obfuscated methods designed to bypass existing safeguards against misuse of model outputs, and that past attempts to stop the activity had not been fully successful.

Weeks later, leading U.S. AI developer Anthropic published its own report detailing similar alleged activity by three major Chinese AI labs: DeepSeek, Moonshot, and MiniMax. Anthropic claimed the three firms had generated over 16 million interactions with Anthropic’s Claude model across roughly 24,000 fraudulent accounts, a clear violation of the company’s terms of service and regional access restrictions. “Distillation can also be used for illicit purposes: competitors can use it to acquire powerful capabilities from other labs in a fraction of the time and at a fraction of the cost that it would take to develop them independently,” the report noted, adding that these attacks follow a consistent pattern: bad actors use proxy services to scale access, build networks of fake accounts to avoid detection, then send massive volumes of structured prompts to extract capabilities and build training datasets, with thousands of nearly identical prompts across coordinated accounts targeting high-value model functions.

The allegations moved to formal congressional scrutiny on April 16, when the House Select Committee on China held a hearing titled “China’s Campaign to Steal America’s AI Edge,” where lawmakers repeated accusations that Chinese firms source high-end Nvidia chips through third countries and use unauthorized distillation to extract proprietary data from U.S. AI models. “Chinese labs are resorting to unauthorized distillation attacks to extract information from our best AI models. Since they don’t have enough AI chips to develop the models on their own, they prefer to simply steal them from their American competitors. Anthropic, OpenAI, and Google have all verified that this is happening,” committee chairman John Moolenaar said, adding that Congress must pass new legislation to block what he called China’s multi-pronged effort to acquire U.S. AI technology through both legal and illegal means.

China has pushed back firmly against these allegations. In response to the White House’s claims of AI IP theft, Chinese Foreign Ministry spokesperson Guo Jiakun called the accusations completely groundless, saying they amount to a deliberate attack on China’s legitimate AI industry development. “We urge the U.S. to respect facts, discard bias, stop its containment of China’s sci-tech development, and choose the course of action conducive to sci-tech exchanges and cooperation between China and the U.S.,” Guo stated.

The escalating focus on AI IP comes as U.S. lawmakers have already ramped up restrictions on Chinese access to advanced semiconductor manufacturing technology. Earlier in April, a bipartisan group of House and Senate lawmakers introduced the Multilateral Alignment of Technology Controls on Hardware (MATCH) Act, which aims to block Chinese chipmakers from purchasing ASML’s deep-ultraviolet immersion lithography systems, critical equipment for producing advanced logic chips. Lawmakers are now expanding that regulatory push beyond semiconductor hardware to target AI model distillation.

These latest developments also come against a shifting backdrop of AI chip policy. Just one day before the OSTP memorandum, on April 22, Commerce Secretary Lutnick confirmed that no Nvidia H200 AI chips have yet been sold to Chinese companies, after Beijing urged domestic tech firms to avoid purchasing the U.S.-made chip, three months after the Trump administration approved H200 exports to China. Industry analysts note that Chinese AI firms still want access to H200 chips but remain wary of potential abrupt shifts in U.S. trade policy that could disrupt supply chains. Against this backdrop, analysts say DeepSeek V4 and Huawei’s recently launched Ascend 950PR AI chip are positioned to become the core of China’s emerging domestic AI ecosystem. Huawei launched the Ascend 950PR in late March, marketing the chip as delivering 2.87 times the performance of Nvidia’s earlier H20 chip and approaching the performance of the H200, with plans to ship roughly 750,000 units of the new chip in 2026.

In private discussions with several Gulf Arab nations, the Trump administration has pushed local authorities to award multi-billion-dollar infrastructure reconstruction contracts to American engineering, manufacturing and construction companies, following the widespread damage the region sustained in Iran’s retaliatory strikes amid the U.S.-Israeli war on the Islamic Republic, multiple U.S. and Arab officials with direct knowledge of the talks confirmed to Middle East Eye.

The Gulf states targeted by the lobbying effort include Kuwait, Bahrain, and the United Arab Emirates – the three countries that bore the brunt of Iranian counterattacks and suffered the worst infrastructure damage. Neighboring Saudi Arabia and Oman, by comparison, reported far less severe damage from Iranian air and drone strikes, the officials added.

In talking points shared with Gulf leaders, U.S. officials have centered their pitch on the deep-rooted economic partnerships between Washington and Gulf monarchies, framing American involvement in reconstruction as a natural extension of those long-standing ties. A senior U.S. official told Middle East Eye that the push to secure reconstruction contracts for domestic firms aligns with the Trump administration’s core “America First” foreign policy doctrine, which prioritizes economic statecraft designed to benefit U.S. commercial interests.

However, one senior Arab official criticized the timing of the lobbying effort, calling it “a little tone-deaf” at a moment when Gulf governments remain on high alert for a resumption of open hostilities, and are increasingly uncertain about long-term U.S. security commitments to the region.

This push is far from a symbolic gesture: independent energy analyst firm Rystad Energy estimates that repair costs for energy-related infrastructure across the Gulf alone could climb as high as $39 billion, a figure that does not include the massive damage sustained within Iran itself. Currently, a fragile ceasefire holds between Washington and Tehran, even as the two sides remain locked in a tense stalemate over control of the Strait of Hormuz, with both enforcing rival blockades on shipping through the critical chokepoint. The Iranian government has calculated that it sustained $270 billion in combined direct and indirect economic damage from the U.S.-Israeli war.

Though most Gulf monarchies publicly opposed the U.S.-Israeli military campaign against Iran, they nonetheless became the primary target of Tehran’s retaliatory strikes. The UAE alone faced at least 2,000 ballistic missiles and drone attacks, according to regional defense sources. The hardest-hit Gulf states are also the most economically vulnerable to Iran’s new control over the Strait of Hormuz, through which roughly a fifth of global oil supplies pass; unlike Saudi Arabia, which operates a pipeline that bypasses the strait via the Red Sea, these nations remain fully reliant on the chokepoint for their energy exports.

While Gulf states collectively hold massive sovereign reserves that can cover reconstruction costs, growing signs indicate they are bracing for a prolonged period of economic downturn as energy exports remain stalled. Kuwait, for example, hosts one of the world’s largest sovereign wealth funds, valued at roughly $1 trillion – on par with the funds held by the UAE and Saudi Arabia, though it generally maintains a far lower public profile. Even so, U.S. Secretary of State Scott Bessent confirmed this week that the UAE and other affected Gulf states have approached Washington to request currency swap lines that would give them access to much-needed U.S. dollars while their energy export revenue remains frozen.

A former senior U.S. official familiar with the negotiations told Middle East Eye that a quid pro quo is already on the table: “I could see the US looking for a trade-off where Gulf states using a swap line commit to US firms for rebuilding.”

Kuwait, located at the northeastern edge of the Persian Gulf, was also heavily damaged by Iranian strikes. The country already hosts the fourth-largest U.S. military presence in the world, and Iranian strikes hit key U.S. facilities including Camp Arifjan and Ali al-Salem Air Base. Beyond military sites, Kuwait International Airport suffered extensive damage, and at least two major national power and water desalination plants were also hit, Reuters reported.

Similarly, Bahrain – a small island kingdom connected only to Saudi Arabia via the King Fahd Causeway – sustained widespread infrastructure damage. The kingdom’s main port, which hosts the U.S. Navy’s Fifth Fleet, was heavily targeted, and critical industrial sites across the country also sustained heavy damage. The Financial Times confirmed that Amazon’s regional cloud computing operations based in Bahrain were knocked offline by the strikes, and Aluminium Bahrain – one of the world’s largest single-site aluminum smelters – was forced to declare force majeure after sustaining critical damage. Bahrain’s national Bapco refinery also issued a force majeure notice following the attacks.

To date, U.S. officials have not yet pushed for contracts to go to any specific American firms, but have made clear their goal of positioning U.S. companies as the top candidates for all major reconstruction work moving forward.

On a Friday briefing at the Pentagon, U.S. Defense Secretary Pete Hegseth doubled down on a controversial demand: the international community should owe President Donald Trump gratitude for launching an unauthorized, unprovoked war against Iran — a conflict that has already upended global energy markets and put millions at risk of imminent food insecurity. The war, which was orchestrated unilaterally by Trump and Israeli Prime Minister Benjamin Netanyahu in late February with no advance consultation or coordination with European allies, has already drained an estimated $60 billion from U.S. taxpayer funds, a cost Hegseth acknowledged but framed as a bold, historic contribution to global security.

Calling the conflict a “bold and dangerous mission” and “a gift to the world courtesy of a bold and historic president,” Hegseth went on to rebuke U.S. allies for refusing to join the military campaign. He argued that alliance commitments are not a one-way street, noting that European nations rely far more heavily on unimpeded access to the Strait of Hormuz, the critical oil and goods chokepoint currently disrupted by the war, than the U.S. does. “They need the Strait of Hormuz much more than we do, and might want to start doing less talking and having less fancy conferences in Europe, and get in a boat. This is much more their fight than ours,” Hegseth told reporters, adding that the U.S. does not count on European support but expects allied nations to step up.

Far from being a boon to the global community, the conflict has already triggered cascading global disruptions that are hitting economies and vulnerable populations hard. According to a Friday report from Barron’s, the war has sparked a widespread global jet fuel shortage that has forced major airlines to slash thousands of scheduled flights. Europe has borne the brunt of the disruption: German flag carrier Lufthansa has announced it will cut 20,000 flights through October, and even major U.S. carriers including Delta have implemented service cuts to offset spiking jet fuel costs.

The risks extend far beyond air travel, with looming threats to global food security that have alarmed senior United Nations officials. The South China Morning Post reported Wednesday that Asian nations are already mobilizing to prepare for widespread food shortages, as the war has cut off global supplies of fertilizer critical for the 2026 northern hemisphere planting season. Compounding this risk, climate scientists have already warned that the year will bring a powerful “super El Niño” event that is projected to reduce rainfall across much of South and Southeast Asia, creating a double blow to crop yields. “It is very concerning because this year is supposed to be a super El Niño, and you are getting into the planting season,” Gnanasekar Thiagarajan, founder of India-based Commtrendz Research, told the outlet. “This is going to be widespread across South and Southeast Asia. There will be dryness everywhere.”

Jorge Moreira da Silva, executive director of the United Nations Office for Project Services (UNOPS), warned this week that the world faces a severe, immediate risk of a full-scale global food crisis if the Strait of Hormuz remains closed to fertilizer shipments. “The planting season has already started, and in most countries in Africa it will end in May,” Moreira da Silva explained. “So, if we don’t get some solution immediately, the crisis will be very significant and severe, particularly for the poorest countries and for the poorest citizens.”

Beyond the global humanitarian and economic fallout of the conflict, the Trump administration and the Pentagon have launched an aggressive crackdown on press freedom covering the war, with Hegseth issuing a new explicit warning to reporters during Friday’s briefing. Hegseth told journalists to “think twice” before publishing stories based on leaked classified information — a standard journalistic practice that has exposed past government abuses including mass surveillance and war crimes. He described such reporting as “incredibly irresponsible and unpatriotic,” and warned that the Pentagon treats leaks “very seriously,” adding a direct rebuke to major outlets including the *New York Times*: “encourage members of the press to think twice about the lives they’re affecting when they publish things in their publications.”

This escalation fits a broader pattern of aggression against press freedom under the current administration amid the Iran conflict. Earlier this month, President Trump publicly stated his administration would seek jail time for journalists who published leaked information about a U.S. fighter jet recently downed over Iran. Just weeks prior, the Pentagon temporarily barred press photographers from on-the-record war briefings after Hegseth’s staff expressed displeasure with unflattering photos of the defense secretary circulating in media coverage.

The Pentagon has also attempted to implement a rule forcing journalists to pledge they will not publish or even solicit any information not explicitly authorized by the department, with violations resulting in permanent revocation of press credentials. A federal judge blocked that policy earlier this month and rebuked the Pentagon for attempting to reimpose the rule after making only insubstantial cosmetic changes. Press freedom advocates warn the policy represents a historic threat to First Amendment protections for investigative reporting. Seth Stern, chief of advocacy at the Freedom of the Press Foundation, wrote in a recent column for *The Intercept* that the administration’s legal arguments go far beyond revoking press access, and would criminalize core work by national security reporters.

“The government argued that although journalists may lawfully ask questions of ‘authorized’ Pentagon personnel, ‘a journalist does solicit the commission of a criminal act, and that solicitation is not protected by the First Amendment, when he or she solicits … non-public information from individuals who are legally obligated not to disclose that information,’” Stern wrote. “The government’s argument would have turned countless Pulitzer-winning national security reporters into criminals.” He added that the Trump administration is expanding on a precedent set by the prior Biden administration, which secured a controversial plea deal with WikiLeaks founder Julian Assange on Espionage Act charges for obtaining and publishing classified government records exposing Iraq War crimes, over repeated objections from First Amendment advocates.

On Friday, a high-profile trial at London’s Woolwich Crown Court heard gripping testimony from six Palestine Action activists charged with criminal damage over an August 2024 break-in at a Bristol-area Elbit Systems facility, with defendants centering their defense on the Israeli arms manufacturer’s development of AI-powered weapons deployed against Palestinian civilians.

The six defendants — Leona Kamio, 30, Charlotte Head, 29, Jordan Devlin, 31, Fatema Rajwani, 21, Zoe Rogers, 22, and Samuel Corner, 23 — all stand accused of vandalism and trespassing stemming from the early-morning incursion into Elbit’s Filton research and development factory, located just outside Bristol in western England.

Taking the witness stand Friday, Rogers, a north London resident with diagnosed autism and ADHD, told the court she and Kamio targeted factory computer systems specifically because of Elbit’s documented work on AI tools that enable more precise targeting of civilians in conflict zones. She explained that the Filton site is not a general manufacturing facility, but a core R&D hub advancing cutting-edge weapons technology for the Israeli government, including AI-driven decision support and targeting systems. Publicly available information on Elbit’s own website confirms the state-owned Israeli firm’s development of world-leading AI-powered decision support systems, built on decades of specialized experience in military simulation and weapons technology.

In his afternoon testimony, Devlin — a product designer from Ballymena, Northern Ireland — echoed Rogers’ framing, calling Elbit’s AI tools the most lethal component of the company’s weapons arsenal. He argued that disrupting the early development stage of these systems would prevent countless future civilian deaths. “If you can disable these AI systems while they are still being refined, you are directly saving lives,” Devlin told the jury. Rogers added that the group’s core goal entering the facility was to disable Elbit’s “killer drones” and save as many civilian lives as possible, telling the court: “I remember destroying weapons used to kill children.”

Rogers told the court she had reviewed public footage of Elbit’s Thor BTOL quadcopter drone, a model the company produces that has been openly deployed in the Gaza Strip, where it is used to drop grenades on civilian populations. Rajwani, Rogers’ co-defendant, confirmed Friday that she specifically targeted Elbit’s quadcopter drones during the incursion, saying: “This is a weapon I knew would kill or injure children. My specific intention was to dismantle drones and other weaponry, and I damaged computer systems along with drone components.”

The trial this week has shown jurors raw on-site footage of the incursion, capturing violent confrontations between the activists, on-site security guards and responding police officers. During the clashes, multiple defendants suffered injuries after being hit with a sledgehammer, struck by police tasers, and sprayed with Pava, a potent synthetic pepper spray authorized for UK law enforcement use. Jurors have also heard details of encounters between defendants and security: footage shown Friday captured on-site guard Angelo Volante shouting at Rajwani and Head to “get on the floor.” Rajwani, a Tanzanian-born British student who works four part-time jobs, told the court the encounter left lasting psychological trauma: “Volante is one of the scariest people I have ever encountered. I still have nightmares about his shouting.”

After her initial arrest for the break-in, Rajwani told the court she was subsequently re-arrested on terrorism offenses. As a visibly Muslim woman of color who grew up in the UK, she described overwhelming fear following the second arrest: “I grew up hearing what terrorism charges mean for people like me. I associate that label with torture and unfair trials. I was terrified I would never get out of custody.” Rajwani added that she brought a GoPro camera to the factory to both document Elbit’s weapons development and live broadcast the action to global audiences.

Rogers, who shared autism and ADHD diagnoses with Corner, told the court she joined Palestine Action after embracing the group’s model of direct action — creating change directly rather than lobbying government officials for reform. Her defense attorney Audrey Cherryl Mogan presented a 2023 Palestine Action document prepared to train activists for the risk of imprisonment, which reads: “Becoming a prisoner for taking action against Israel’s arms trade is proof of causing significant costs to the arms industry and its protector, the imperial British state. In a neutered, pacified society that tolerates a business model built on the genocide of Palestinian people, taking action is not only crucial, but a rare act of meaningful solidarity.” Rogers confirmed she shared this core belief, though she told the court she never wanted to be imprisoned, and has already been unable to resume her university studies while detained. She also criticized the UK’s prison system as a profit-driven enterprise.

In closing testimony Friday, Devlin shared his personal background growing up Catholic in the majority-Protestant town of Ballymena, noting that a British soldier saved his grandfather’s life after he was nearly beaten to death by loyalist paramilitaries. The court was also told that Devlin, a working product designer, created a statue that won singer Sam Fender the 2023 Mercury Prize. Devlin told the jury he has seen senseless violence up close throughout his life, saying: “I grew up surrounded by violence, and it has always seemed senseless to me that humans inflict this harm on each other. There is no reason to accept extreme violence as inevitable, when we can stop it at its source.” He added that Palestine Action’s core goal is to force Elbit to cease all operations in the UK: “It is an absolutely horrific company, and I cannot understand why it is still allowed to operate here.”

On Thursday, Corner testified about his confrontation with Sergeant Kate Evans, who was struck twice during the incursion. Corner told the jury he had already been incapacitated by Pava spray, and mistook Evans for an aggressive security guard attacking a fellow activist. The trial is ongoing.