During the European Central Bank's annual forum in Sintra, global monetary authorities examined the systemic risks and operational shifts precipitated by the integration of artificial intelligence within the financial sector.

在歐洲央行於辛特拉舉行的年度論壇中，全球貨幣權威研究了將人工智能整合至金融部門所引起的系統性風險與操作轉變。

The discourse centered on the potential for AI to induce significant volatility through the acceleration of asset bubbles. The Bank for International Settlements noted that current investment trajectories mirror historical speculative episodes, such as the 19th-century railway mania, suggesting a risk of substantial near-term corrections. Furthermore, Professor Itay Goldstein highlighted the capacity for algorithmic coordination to facilitate price manipulation, which could undermine market integrity.

討論的焦點在於 AI 是否有可能透過加速資產泡沫而導致嚴重波動。國際結算銀行指出，目前的投資趨勢與歷史上的投機事件相似，例如 19 世紀的鐵路狂熱，顯示短期內有大幅度修正的風險。此外，Itay Goldstein 教授強調了演算法協調可能會便利價格操縱，從而損害市場完整性。

Institutional stakeholders expressed concern regarding the transition from generative AI to 'agentic' systems. Deputy Governor Sarah Breeden of the Bank of England posited that the autonomy of these agents—capable of executing multi-step transactions without human intervention—renders traditional 'human-in-the-loop' oversight obsolete. This shift may precipitate herd behavior, wherein synchronized algorithmic responses to market signals amplify systemic instability, necessitating the implementation of regulatory 'kill switches' or circuit breakers.

機構持份者對從生成式 AI 轉向「代理」系統表示擔憂。英格蘭銀行副行長 Sarah Breeden 認為，這些代理具有自主權——能夠在無需人類干預的情況下執行多步驟交易——使得傳統的「人類參與」監管變得過時。這種轉變可能會引起羊群效應，即演算法對市場訊號的同步反應會放大系統不穩定性，因此有必要實施監管「緊急停止開關」或熔斷機制。

From a supervisory perspective, the IMF's Tobias Adrian identified a critical deficit in 'explainability,' noting that the 'black box' nature of AI-driven credit decisions complicates regulatory auditing. Additionally, the potential for a widening digital divide between affluent and marginalized economic entities was discussed, alongside the escalating costs of cybersecurity. To mitigate these vulnerabilities, the Bank of England proposed the conceptualization of insurance frameworks to maintain essential functions during systemic disruptions.

從監管角度看，IMF 的 Tobias Adrian 指出「可解釋性」嚴重不足，提到 AI 驅動的信貸決定具有「黑盒」特性，使得監管審計變得複雜。此外，與會者亦討論了富裕經濟實體與邊緣實體之間數位鴻溝擴大的可能性，以及網路安全成本上升的問題。為了降低這些漏洞，英格蘭銀行建議構思一套保險框架，以便在系統性中斷期間維持基本功能。

Finally, the macroeconomic implications remain bifurcated. A scenario of excessive efficiency could lead to mass labor displacement and a subsequent contraction in aggregate demand, while a failure to realize projected productivity gains could result in a severe capital devaluation, reminiscent of the dotcom era.

最後，宏觀經濟影響仍然分歧。一種是效率過高的情境，可能會導致大規模勞動力流失，進而令總需求縮減；而另一種是若未能實現預期的生產力提升，則可能會導致嚴重的資本貶值，令人聯想到 dotcom 時代。

Global financial regulators are currently evaluating the necessity of new guardrails to manage the transition toward autonomous AI agents and the associated risks to market stability.

全球金融監管機構目前正在評估是否有必要建立新的護欄，以管理向自主 AI 代理過渡的過程，以及對市場穩定造成的相關風險。