The artificial intelligence industry is currently facing increased government regulation, growing competition between companies, and serious concerns regarding the accuracy of AI-generated information.

人工智慧產業目前面臨政府監管增加、企業間競爭加劇，以及對於 AI 生成資訊準確性的嚴重疑慮。

Public tension is rising, as seen when students at Stanford University protested against Google CEO Sundar Pichai. The protests focused on Project Nimbus, a $1.2 billion cloud contract with the Israeli government. At the same time, the industry is experiencing instability. For example, the government ordered a suspension of Anthropic's Fable 5 and Mythos 5 models after Amazon researchers discovered security flaws that could lead to cyberattacks. Consequently, many are questioning whether safety standards for advanced AI models are consistent.

公眾緊張局勢正在升級，例如史丹佛大學的學生對 Google 執行長 Sundar Pichai 進行抗議。抗議焦點在於 Project Nimbus，這是一份與以色列政府價值 12 億美元的雲端合約。與此同時，該產業正經歷不穩定期。例如，在亞馬遜研究人員發現可能導致網路攻擊的安全漏洞後，政府下令暫停使用 Anthropic 的 Fable 5 和 Mythos 5 模型。因此，許多人質疑先進 AI 模型的安全標準是否一致。

Furthermore, there is a significant lack of reliability in professional AI content. An analysis by GPTZero of a KPMG report found that about 50% of the claims were incorrectly attributed, and only five out of 45 citations were actually real. This problem of 'hallucinations' also appears in corporate marketing, such as the false claim that Emirates' 'Sara' assistant could change flight bookings.

此外，專業 AI 內容嚴重缺乏可靠性。GPTZero 對一份 KPMG 報告的分析發現，約 50% 的主張引用錯誤，且 45 個引用中僅有 5 個是真實的。這種「幻覺」問題也出現在企業行銷中，例如阿聯酋航空的「Sara」助手聲稱能更改航班預訂，但實際上並非如此。

Internally, Meta has started the Model Capability Initiative (MCI), which tracks employee keystrokes and mouse clicks to improve AI models. While leadership defends this as a way to collect high-quality data, it raises privacy concerns. Meanwhile, a 'cold war' exists between Anthropic and OpenAI. Anthropic CEO Dario Amodei emphasized that a group of 'trustworthy' companies should set industry standards to ensure safety. These corporate struggles are happening while G7 leaders and tech executives meet to discuss how to protect human dignity and maintain control over AI systems.

在內部方面，Meta 啟動了「模型能力計畫」(MCI)，透過追蹤員工的按鍵和滑鼠點擊來改良 AI 模型。雖然領導層將其辯稱為收集高質量數據的方式，但卻引發了隱私疑慮。同時，Anthropic 與 OpenAI 之間存在一場「冷戰」。Anthropic 執行長 Dario Amodei 強調，應由一群「值得信賴」的公司制定行業標準以確保安全。當 G7 領袖與科技高層會面討論如何保護人類尊嚴並維持對 AI 系統的控制時，這些企業鬥爭正同步發生。

The AI sector remains divided, showing a clear gap between rapid technological growth and the difficult effort to create strong ethical and operational rules.

AI 產業仍然分歧，顯示出快速的技術成長與建立強而有力的倫理及操作規範之艱難努力之間，存在明顯差距。