The United States is currently facing institutional tension due to conflicting AI management strategies between the federal government and state leaders, particularly in California.

由於聯邦政府與州領袖（特別是加州）在 AI 管理策略上存在衝突，美國目前正面臨體制上的緊張局勢。

The federal government is currently struggling to balance public authority with private innovation. To protect national security, the administration has used export controls to limit advanced AI models. For example, the Pentagon labeled the company Anthropic as a supply chain risk and restricted its Fable 5 model. These strict measures were caused by concerns that the Mythos model could identify software vulnerabilities, marking a shift from a hands-off approach to much tighter oversight.

聯邦政府目前正努力在公共權力與私人創新之間取得平衡。為了保障國家安全，行政部門利用出口管制來限制進階 AI 模型。例如，五角大廈將 Anthropic 公司列為供應鏈風險，並限制了其 Fable 5 模型。這些嚴格措施是由於擔心 Mythos 模型可能會識別出軟體漏洞，標誌著從放任不管轉向更嚴格的監管。

At the same time, California has taken a different path by working more closely with the private sector. Governor Gavin Newsom signed an agreement with Anthropic to use the Claude AI system in all state and local agencies at a 50% discount. This move aims to make government work more efficient, even though the federal government continues to argue with the company over issues like mass surveillance and autonomous weapons.

與此同時，加州採取了不同的路徑，與私部門更緊密地合作。州長 Gavin Newsom 與 Anthropic 簽署協議，讓所有州和地方機構以 50% 的折扣使用 Claude AI 系統。此舉旨在提高政府工作效率，儘管聯邦政府仍就大規模監控和自動武器等議題與該公司爭論不休。

Furthermore, public opinion shows that people from both political parties want mandatory safety reviews. According to the AI Policy Institute, most voters prefer government-mandated safety standards rather than letting companies regulate themselves. This creates a gap between the administration's current policies and what the public expects. Meanwhile, China's Zhipu AI has reached a similar level of technical ability as the Mythos model. Because the US is facing internal conflicts between different levels of government, it risks losing its competitive edge in the global AI race.

此外，公眾輿論顯示，兩黨民眾都希望建立強制性的安全審查。根據 AI 政策研究所，大多數選民偏好政府強制的安全標準，而非讓公司自行監管。這造成了行政部門現行政策與公眾期望之間的差距。同時，中國的智譜 AI 已達到與 Mythos 模型相近的技術能力。由於美國面臨不同層級政府之間的內部衝突，恐將在全球 AI 競賽中失去競爭優勢。

The US AI sector remains divided, as federal security goals clash with state-level adoption and a growing public demand for official safety regulations.

美國 AI 領域依然分歧，因為聯邦安全目標與州級採納，以及公眾對官方安全法規日益增加的需求之間存在衝突。