This paper looks at the implications for monetary policy of the widespread adoption of artificial intelligence and machine learning, which is sometimes called the “fourth industrial revolution.” The paper reviews experiences from the previous three industrial revolutions, developing a template of shared characteristics:new technology displaces workers;

investor hype linked to the new technology leads to financial excesses; new types of jobs are created; productivity and potential output rise; prices and inflation fall; and real debt burdens increase, which can provoke crises when asset prices crash.

The experience of the Federal Reserve during 1995–2006 is particularly instructive. The paper uses the Bank of Canada’s main structural model, ToTEM (Terms-of-Trade Economic Model), to replicate that experience and consider options for monetary policy. Under a Taylor rule, monetary policy may allow growth to run as long as inflation remains subdued, easing the burden of adjustment on those workers directly affected by the new technology, while macroprudential policies help check financial excesses. This argues for a family of Taylor rules enhanced by the addition of financial stability considerations.

本文着眼于人工智能和机器学习的广泛采用对货币政策的影响，人工智能和机器学习有时被称为“第四次工业革命”。本文回顾了前三次工业革命的经验，开发了具有共同特征的模板：新技术取代了工人；

与新技术有关的投资者炒作导致财务过剩；创建新的工作类型；生产率和潜在产出增加；价格和通胀下降；实际债务负担增加，这可能在资产价格暴跌时引发危机。

美联储在1995年至2006年期间的经验特别具有启发性。本文使用加拿大银行的主要结构模型TOTEM（贸易条件经济模型）来复制这种经验并考虑货币政策的选择。根据泰勒（Taylor）规则，只要通货膨胀保持在低位，货币政策就可以允许增长持续下去，从而减轻了直接受新技术影响的工人的调整负担，而宏观审慎政策则有助于遏制金融过剩。这为通过增加财务稳定性考虑而增强的泰勒规则系列提供了依据。