Citigroup intends to provide more than 40,000 of its programmers with access to generative artificial intelligence amid Wall Street’s growing interest in advanced technology that is projected to surpass human mental abilities in the foreseeable future.

As part of the pilot program, which was known only within the corporate space of the specified financial institution, the bank allowed approximately 250 of its developers to conduct experiments using generative machine intelligence. The results of this project are unknown, but the fact that next year the lender intends to scale the program for the majority of its programmers indicates certain successes and signals that Citi sees significant prospects for itself in applying artificial intelligence.

This financial institution and its competitors began experimenting with AI after a chatbot based on an advanced technology called ChatGPT was presented last year, which demonstrated impressive capabilities. Generative artificial intelligence can write texts based on user requests.

In the banking sector, the opinion that the use of machine intelligence is a way to increase the level of efficiency of work processes has been spreading more and more actively lately. For example, Citi applied generative artificial intelligence to read 1089 pages of new rules on capital for financial institutions in the United States, which were published by federal regulators.

The mentioned bank’s risk management and compliance team used AI to calculate how much capital should be prepared to create protection against future losses. Generative artificial intelligence divided the proposals on this issue into parts and drew up the main conclusions.

For Citi, the appearance of ChatGPT has become something like a signal to start using AI. This year, the financial institution has formed two task forces to explore options for applying artificial intelligence. This was announced by Stuart Riley, the bank’s co-head of information technology, who oversees the relevant work, which, according to him, concerns all units of the lender. He also noted that some solutions help to cope with tasks that relate to the so-called daily routine, while other AI use cases are designed to carry out more complex activities.

Many employees in the banking sector are concerned that they may lose their jobs due to machine intelligence. Stuart Riley does not share this vision of the future. According to him, artificial intelligence needs human supervision, regardless of what kind of work it does. In this context, Mr. Riley noted that AI tools help developers create code faster, but do not replace them. Separately, he stressed that the bank uses artificial intelligence to enhance the capabilities of employees.

Citi is also currently exploring ways to upgrade its systems with advanced technology. To update existing systems, a financial institution needs to change the programming language, which AI will help to do.

The bank also considers artificial intelligence as a tool for analyzing documentation and generating content. AI can potentially analyze revenue and other publicly available documents.

Another option for using advanced technology that a financial institution is considering is scanning data arrays for errors and anomalies. Stuart Riley says that this AI-applying format can be useful for making payment transactions following credit agreements. According to him, this is a way to complete the cycle, speed up processes, and increase the end-to-end accuracy of recording operations, up to a legal contract.

Citi is also exploring the prospects of using large language models to analyze the legislation of the countries in which the bank operates. In this case, the main goal is to ensure that the activities of a financial institution comply with regional regulatory legal acts.

As we have reported earlier, Bank of America Uses AI to Train New Hires.