Abstract: Mr. Tobias Adrian believes that AI has brought positive changes to the financial sector—algorithmbased financial transactions have increased market efficiency and reduced financing costs across sectors; using AI to improve credit allocation enhances financial inclusiveness. Theoretically, with the emergence of general AI in the future, the financial sector may also achieve autonomous decision-making by machines.
However, from the perspective of financial stability, AI increases the difficulty of risk management at the corporate level, and can also expand the scale of attacks (against financial institutions), increasing the efficiency of attacks, with cyber risks becoming a potential macro critical risk. Meanwhile, although generative AI may be proficient in understanding cross-sectional risks, it is unclear whether it can capture general equilibrium effects and how this would impact financial cycles.
On macro stability, Prof. Jason Furman believes that in the short term, the application of AI will not increase output but will only increase labor input, thus overall productivity is declining. This also means that managing inflation has not become easier in the short term, and may in fact increase, requiring higher interest rates to stabilize the macro-economy. However, in the long run, the development of AI allows for significant growth in productivity, which will translate into increased income and higher interest rates.
Regarding the commonly concerned issue of labor replacement, both Adrian and Furman are optimistic because historical experience shows that with technological progress, new types of jobs will emerge, and people’s overall income levels will rise, leading to more service industry positions. Moreover, technology can only replace some jobs, not all. However, Furman also emphasizes the risk of increasing inequality, such as people lowering their wages to compete with robots. He believes that wide-ranging measures must be taken to address this challenge, one focus being the investment of more educational resources.
Regarding the regulatory challenges of artificial intelligence, Adrian believes that a major challenge currently facing financial regulators is that the data needed to understand these risks are not necessarily the same as those needed in the past, and where and what kind of data to collect may differ from before. Therefore, increasing transparency may be the top priority. Meanwhile, regulators need to try to use AI to provide information for regulatory visibility and system stability judgment.
Furman believes that waiting for regulatory measures to catch up with the development of AI is not a wisemove. Slowing down the deployment of AI also carries many risks. When considering AI regulation, the first priority is to balance benefits and risks. Second, attention should be paid to AI biases and not to view it as omnipotent; third, regulation should not become a moat to protect existing businesses. Finally, he points out that many solutions to AI issues are not related to regulating AI but involve independent strategies concerning the labor market, tax systems, etc.
