Moody's has warned that the rush to adopt artificial intelligence is making big banks dependent on a small group of Silicon Valley companies, exposing them to widespread outages and potential price increases.
Potential benefits and risks
The rating agency said integrating AI into daily operations could eventually lower costs and boost revenues across the financial sector, but it requires significant investment. With many competitors pursuing the same goal, these benefits may be eroded.
AI also introduces greater risks around data privacy, cybersecurity, fraud, and deposit flight, as well as over-reliance on a few tech providers, Moody's cautioned.
Current adoption and concerns
More than 75% of City firms now use AI, according to a UK Treasury select committee report from January, with insurers and international banks leading adoption. They primarily use it for administrative tasks and core operations like claims processing and credit assessments.
Moody's report highlighted that reliance on a small set of foundation AI models and cloud providers could create systemic dependency. A single provider outage could spread quickly across customers and sectors, prompting regulators to focus on operational resilience and third-party concentration.
Vendor dependence and mitigation
The AI race also creates vendor dependence risk, where dominant providers could control AI service prices. This may emerge as generative AI companies like OpenAI and Anthropic face pressure to generate profits.
Despite these risks, financial firms retain control over key assets like proprietary data. Many banks and insurers have experience negotiating tech contracts and may use open-source models or partnerships to reduce dependency.
Industry response
Lloyds Banking Group's CEO, Charlie Nunn, recently announced a £13bn strategy involving AI to attract business, improve efficiency, and increase shareholder payouts. This includes £2bn in cost cuts, which Nunn said would affect staff and require reskilling.
Moody's acknowledged potential job losses, estimating a 20% chance that by 2030 AI could perform the work of a solid mid-level employee. For banks, AI may also enable customers to switch accounts for better rates, risking large cash movements, making depositor trust and funding stability critical.
