Deep Dive into the Storm Behind Anthropic's Account Suspensions: The Security Religion, AI Civil War, and Claude's Predicament Amid US-China Decoupling

In April 2026, employees of an American agricultural technology company called Agricultural Technology Company started their day as usual, opening their computers to use Claude Code for writing code, data analysis, and supply chain analysis. Without any warning, they found that 110 employee accounts had all been suspended. The company's network administrator received an email from Anthropic stating: Activity violating our usage policy has been detected, and your account has been suspended.

Although the accounts were collectively suspended, the backend API continued to function normally and charges were still deducted. The company's network administrator even received payment reminder messages. Subsequently, the company's management sent appeal emails and contacted Anthropic, but ultimately made no progress. Claude Code's shutdown directly brought the entire team's work to a standstill.

Around the same time, on Chinese internet platforms like V2EX, Zhihu, and Juejin, user complaints about Claude were everywhere: someone had just topped up their Max subscription and their account was banned within seconds before they could even use it; someone used a virtual card to pay, and the system immediately flagged it as "account violation" upon successful payment; someone logged in using a third-party tool and was blacklisted by the system, getting four accounts banned in three months without a single successful appeal.

In fact, ever since Anthropic entered the market with its flagship product, Claude Code, and ascended to the top tier, it has earned a reputation as the "ban-happy" champion.

According to risk control data for the second half of 2025, published by Anthropic's Transparency Hub in January 2026, the platform banned a total of 1.45 million accounts in just six months. Of these, 52,000 appeals were filed, but only 1,700 were successful. This means the appeal success rate was merely 3.3%.

Source: https://www.anthropic.com/transparency

In other words, out of 100 users who feel particularly wronged by a ban, only about 3 are likely to get their accounts back. The remaining 97 have to accept their bad luck.

This also indicates that Anthropic does not operate on the principle we might understand—first clarifying the facts, then applying penalties according to rules. Instead, it leans towards preventive enforcement. Its core goal is to intercept risks in their infancy through high-coverage screening, preferring to wrongly punish 1,000 rather than let one slip through.

In contrast, its counterparts ChatGPT and Google Gemini are relatively more lenient.

ChatGPT is much more tolerant of third-party tools and edge-case prompts, with relatively relaxed ban policies.

Gemini, even when it occasionally tightens risk controls, rarely issues warnings, holds accounts collectively responsible, or engages in mass banning.

Only Anthropic treats account suspensions as routine, especially with Claude Code, which has become a major hotspot for bans.

So why is Anthropic's user policy so strict? I believe the reasons are complex.

This involves not only the lifelong obsessions of founder Dario Amodei, the factional split at OpenAI, the power struggles in Silicon Valley capital, and the civil war between the AI safety and acceleration camps in the US, but also the geopolitical chess game of US-China AI decoupling—a grand game hidden behind the code concerning the future control of AI and global technological barriers.

In this article, let's deconstruct this layer by layer.

01. Dario Amodei's Obsession

The root of Anthropic's stringent risk control lies hidden in the life trajectory of founder Dario Amodei. Every choice he made, every obsession he held, has ultimately been transformed into Anthropic's "zero-tolerance" iron law, and likewise into the account suspension emails for countless users.

Recent official portrait of Dario Amodei Source: https://fortune.com

In 1983, Dario Amodei was born into an ordinary immigrant family in San Francisco. His father was an Italian-American leather craftsman who lived his whole life by his craft, with a stubborn personality that valued right and wrong above all else.

His mother was Jewish, working on library renovation projects. She was meticulous in her work and instilled in Dario from a young age the philosophy that "responsibility trumps everything."

Growing up in this family atmosphere, Dario developed a personality characterized by sticking to his principles and maintaining firm bottom lines, allowing no room for ambiguity or compromise.

Young Dario showed the traits of a scientific prodigy. He didn't enjoy crowds or socializing, channeling all his energy into mathematics and physics. When textbook knowledge wasn't enough, he'd immerse himself in the library, devouring various profound theoretical works. At that time, his greatest dream was to become a theoretical physicist and explore the ultimate mysteries of the universe.

In 2006, Dario's father fell ill with a rare and difficult disease that defied cure despite consulting numerous renowned doctors, eventually passing away. His father's death dealt a devastating blow to the 20-year-old Dario, completely overturning his worldview.

Watching his father suffer from illness and facing the dilemma that medicine couldn't solve, he suddenly realized that abstract theoretical physics couldn't save the people before him, couldn't help ordinary people afflicted by diseases.

Consequently, he resolutely abandoned the theoretical physics he had studied for years and switched to biophysics, determined to "use science to cure human diseases," embedding the principle of "controlling uncontrollable risks" deep within his being.

This obsession runs through his entire career:

He first entered the California Institute of Technology for his undergraduate studies, then transferred to Stanford University to complete a Bachelor's degree in Physics. Subsequently, he was admitted to Princeton University for a PhD in Biophysics, becoming a Hertz Foundation Fellow, specializing in the relationship between biomolecular structure and disease. After his doctorate, he pursued postdoctoral research at Stanford University School of Medicine, continuing to delve into the biomedical field, trying to find ways to combat rare diseases.

It wasn't until 2014 that Andrew Ng extended an olive branch, inviting him to join Baidu's US lab, marking his first exposure to artificial intelligence.

At that time, AI development was in its very early stages, mainly used for image recognition and speech synthesis. However, Dario keenly realized that AI could not only change lives but also become a super tool for countering risks and saving humanity. But this premise required it to be strictly controlled and not spiral out of control.

After leaving Baidu, he joined Google Brain as a Senior Research Scientist, delving deep into the field of deep learning, focusing on AI safety—specifically, how to make AI obedient and prevent it from harming humans.

It was also during this phase that he began contemplating how to truly embed human values into AI's core logic, rather than simply applying post-hoc filters.

In 2016, shortly after OpenAI was founded, it attracted top global AI talent with its slogan of "open source, non-profit, advancing AI to benefit humanity." Dario was drawn to OpenAI's philosophy and joined. Leveraging his top-tier technical skills, he rose from head of the AI safety team to Research Director, and then to Vice President of Research, participating fully in the development processes of GPT-2 and GPT-3.

Photo from Dario Amodei's early career period (OpenAI/Google Brain era, circa 2018-2021) Source: bigtechnology

During this time, he was also a co-inventor of RLHF (Reinforcement Learning from Human Feedback). Simply put, this technology uses human feedback to correct AI outputs, making AI more aligned with human values. It later became a safety patch for the entire AI industry. At that time, Dario was full of passion for making AI safe in practice and application, not realizing that his ideals would soon be shattered by reality.

OpenAI's Internal Strife: The Factional Split Between the Safety and Acceleration Camps

Many people know that Dario Amodei left OpenAI with his team in 2021 to found Anthropic, but few know that behind this "defection" lay a years-long conflict of ideas, power struggles, and a profound sense of "betrayal" for Dario.

In its early days, OpenAI indeed adhered to the principles of "non-profit, safety-first." Elon Musk was an early investor who always emphasized that AI safety was paramount. However, as time passed, especially after Sam Altman became CEO of OpenAI, the company's development direction began to shift dramatically.

Sam Altman is a classic "accelerationist." He believes that the pace of AI development must match the speed of the times: first, build larger, more powerful models to seize market opportunities and achieve commercialization, and only then address safety issues.

Symbolic image of the OpenAI and Anthropic factional split (Sam Altman vs Dario collage) Source: wsj.com

Under his leadership, OpenAI began to downplay its "non-profit" identity, actively seeking commercial partnerships, and even proactively moving closer to Microsoft to secure more funding and computing power—all to rapidly iterate the GPT series models and capture a larger share of the AI market.

But all of this was unacceptable to Dario Amodei.

In his eyes, AI was not merely a tool for capturing market share, but a "civilization-level force that could either cure humanity or destroy it." Without first solving safety issues and ensuring AI alignment with humans, the consequences of a model spiraling out of control would be catastrophic.

He repeatedly proposed internally to slow down the pace of model iteration, strengthen safety testing, and prioritize "alignment first." Yet his voice became increasingly marginalized.

In fact, ideological differences were just the surface. The deeper conflict lay in power struggles and credit attribution.

According to a 2026 in-depth report from The Wall Street Journal, Dario Amodei made core contributions to the development of GPT-3—specifically, the implementation of RLHF technology was led by him. However, in public communications, his contributions were significantly downplayed. Sam Altman's team tended to emphasize the "scale and capabilities of the model," neglecting the safety technology spearheaded by Dario.

Adding to Dario's disillusionment, after Musk left OpenAI due to ideological differences, company leadership fell entirely into Sam Altman's hands. The safety team's budget was slashed, many core safety R&D projects were suspended, and some senior executives openly stated: "Safety issues can be put on hold. Let's focus on commercialization first. Once we have money, we can come back and solve safety problems."

Dario knew he could no longer realize his dream of "safely deploying AI" within OpenAI. Recalling this experience later on Lex Fridman's podcast, his tone was calm but carried a hint of finality: "Arguing with others about core vision is an incredibly unproductive thing. Instead of wasting time, it's better to build your own team and realize your own ideals."

In early 2021, AI prodigy Dario made a decision that shocked Silicon Valley: he led his sister Daniela Amodei (now President of Anthropic), along with OpenAI's core safety team and key research personnel, in a mass exodus.

Photo of Dario Amodei with his sister Daniela Amodei Source: Fortune

This exodus is considered a thorough repudiation of OpenAI's accelerationism and a steadfast defense of the safety-first philosophy.

At that time, OpenAI officially issued a courteous statement congratulating Dario's team on their new journey. But privately, the rift between the two sides was irreparable.

What Dario took with him was not just top-tier talent, but also OpenAI's most core safety technologies and philosophy, which ultimately shaped Anthropic. After Dario's departure, OpenAI went entirely down the path of commercial acceleration, moving further and further away from Dario's original vision.

Source: https://openai.com/index/organizational-update/

Anthropic's "Safety Religion"

In February 2021, Dario Amodei formally founded Anthropic, positioning it as a "public-benefit corporation." This means the company's core objective is not profit maximization but "advancing the safe and controllable development of AI for the benefit of humanity."

The obsession with "controlling risk" born from his father's death, and the "safety first" principle forged during his departure from OpenAI, eventually became Anthropic's core system—a "safety religion" etched into the company's DNA.

From its inception, Anthropic established a core invention called Constitutional AI. This was the crystallization of Dario's years of thinking about "AI safety" and the key differentiator from OpenAI and Google Gemini.

Schematic of Constitutional AI Source: Aashka Patel

Constitutional AI does not follow OpenAI's approach of using RLHF as a "post-hoc patch." Instead, it implants a "constitution" into the AI's underlying training process. This constitution integrates the UN Universal Declaration of Human Rights, common ethical principles of humanity, and Anthropic's own safety principles. It requires the AI to "self-review" and "self-criticize" before generating any output or executing any command, ensuring that outputs align with human values and do not produce any dangerous content.

Dario personally wrote two long pieces, "Machines of Loving Grace" and "The Adolescence of Technology," detailing his AI vision:

He believes AI is like an adolescent—possessing immense potential but also full of uncertainty. It is necessary to establish rules and build defenses in advance to prevent it from going astray. Constitutional AI is the "rule book" for AI, playing the role of a defensive line.

This safety religion is reflected not only in model training but is directly transmitted to every product and risk control policy at Anthropic. Claude Code's high-privilege design, combined with prompt injection probes and conversation classifiers, ensures the AI performs an extra layer of self-check before executing commands. The risk control logic of preventive enforcement—preferring to wrongly punish innocents rather than miss any suspicious behavior—aims to nip risks in the bud.

The 2026 incident where Anthropic stood up to the US Department of Defense best exemplifies this "safety fundamentalism." This event not only shocked Silicon Valley but also showed the world Dario Amodei's determination to sacrifice profit rather than abandon safety.

In early 2026, the US Department of Defense asked Anthropic to remove two major safety guardrails from Claude:

First, prohibit Claude from being used for "mass surveillance of US citizens."

Second, prohibit Claude from being used for the development and deployment of "fully autonomous lethal weapons."

The DoD promised that if Anthropic complied, it would sign a military contract worth $200 million and provide substantial computing power support.

At that time, Anthropic was under significant computing power and financial pressure. The $200 million military contract could have alleviated the company's immediate difficulties.

But Dario Amodei refused outright.

He issued a public statement with a firm tone: "We cannot violate our conscience to develop technology that may harm humans or infringe on human rights. Claude's safety guardrails are our bottom line, and we will not compromise."

His refusal infuriated the US Department of Defense. Under the Trump administration's direction, the DoD placed Anthropic on a "supply chain risk" blacklist—the first time in US history a domestic AI company was added to this list. This meant all US defense contractors were prohibited from using Anthropic's products and services. Furthermore, the DoD threatened to invoke the Defense Production Act to force Anthropic to remove its safety guardrails.

Facing pressure from the state apparatus, Dario Amodei directly sued the US Department of Defense, arguing that this action constituted "retaliatory punishment against Anthropic" and violated US laws and values. Although the appeals court later denied Anthropic's request for a temporary injunction, Dario never compromised. Even though the company lost a huge contract and was ostracized by the entire US military-industrial complex, he steadfastly held onto his "safety bottom line."

Seeing this, we