JPMorgan Chase, the largest U.S. bank, is adopting OpenAI's AI models for software development and cybersecurity. As the bank expands AI from a tool for routine work to core IT system development and security tasks, changes are expected in competition among financial firms over AI use.

American Banker reported on Nov. 10 that JPMorgan Chase signed an agreement with OpenAI to adopt AI technology. OpenAI President Greg Brockman (그레그 브록먼) disclosed the partnership and said the companies would expand AI use in coding and cybersecurity.

JPMorgan Chase plans to use OpenAI's coding agent Codex to support software development processes such as writing and reviewing code, fixing errors and deployment. It will also apply OpenAI models to cybersecurity work.

The agreement is drawing attention because JPMorgan Chase rarely publicly discloses partnerships with external technology companies. Analysts say the technology adoption strategy of the largest U.S. bank could also affect the direction of AI investment by other financial institutions.

JPMorgan Chase already uses several AI services for work. Through its in-house platform LLM Suite, it provides OpenAI's GPT-4 and Anthropic's Claude, among others, and its development organization has used coding tools such as GitHub Copilot.

It will, however, maintain a strategy of not becoming dependent on any one company's technology. The bank said it would continue to use multiple AI providers. It did not disclose the contract amount or which OpenAI models it will introduce beyond Codex.

The bank is also increasing AI investment. JPMorgan Chase Chief Financial Officer Jeremy Barnum (제러미 바넘) estimated annual AI-related spending at about $700 million at last month's earnings announcement. He said productivity improvements from AI investment have not yet been clearly confirmed in quarterly results.

JPMorgan Chase is also using open-source and open-weight models that it runs in its own data centres. The strategy is to reduce vendor dependence and manage costs by choosing from various models.

In the financial sector, attention is focused in particular on cybersecurity applications. That is because protecting customers' financial information and core systems requires a higher level of trust and control than typical assistance with document writing.

The Office of the Comptroller of the Currency has also recently issued supervisory guidance to assess banks' AI use in connection with existing risk management frameworks. As financial institutions expand AI use, security and operational risk management are emerging as key tasks along with productivity.