A recent report (June 2024) from Citigroup researchers states that the finance sector will be “at the forefront” of changes driven by artificial intelligence. Banking jobs are identified as most at risk for AI-driven displacement, yet the adoption of AI in finance will likely be slow due to regulatory challenges and other factors.
AI has long been anticipated to profoundly change jobs across all industries. However, Citi's report emphasizes that “finance will be at the forefront of these changes.” The report underscores that the appearance and operation of banks and financial firms in the mid-2020s will differ significantly from those in the mid-1980s or mid-1940s. AI, the report suggests, will accelerate this transformation.
General-purpose technologies (GPTs) such as AI create new opportunities for innovation and can enhance quality of life. However, they also disrupt existing practices, leading to short-term displacement. According to Citi, data from Accenture Research and the World Economic Forum indicates that approximately 67% of banking jobs have a higher potential to be automated or augmented by AI, putting them at the highest risk of AI-led job displacement. Despite this, Citi suggests that a decline in headcount may be counterbalanced by an increase in roles related to AI compliance, ethics, and governance.
The report does highlight a positive aspect: Citi estimates that the global banking sector's profit pool for 2023 could rise by 9% or $170 billion due to AI adoption, increasing from just over $1.7 trillion to nearly $2 trillion.
However, AI adoption in finance will be slow. The Citi researchers attribute this to the highly regulated nature of the sector and the lack of globally aligned rules. The evolving regulatory landscape poses a challenge, with countries moving at different speeds and taking varied approaches to regulation.
Shameek Kundu, head of financial services and chief strategy officer at TruEra, echoes this sentiment in the report. He describes traditional AI adoption in financial services as “widespread, shallow, and inconsequential.” Kundu notes that while many enterprises experiment with AI across various use cases, there is a limited scale of AI adoption and a minimal perceived impact of AI failures on critical business operations.
Citing a 2022 Bank of England survey, Kundu points out that “72% of firms reported using or developing machine learning applications.” However, the median number of ML applications for mainstream UK financial institutions is only 20-30, with less than 20% of these AI use cases being critical to business operations. You can read the full report here.