The news: ChatGPT has dominated headlines. But major banks like JPMorgan, Bank of America, Citi, Goldman Sachs, and Wells Fargo recently banned using it for business purposes, per Bloomberg.
As the industry figures out how to use generative artificial intelligence (AI) safely and effectively, we dig into what the technology could mean for banking in our report: ChatGPT and Generative AI in Banking.
Four banking verticals will feel the greatest impact.
- Retail banking and wealth: Firms will use generative AI to train models behind know-your-customer (KYC) processes for account opening. It will also power natural language models to fine-tune virtual assistants.
- SMB banking: The technology will efficiently comb through non-numerical loan application data like business plans.
- Commercial banking: Generative AI will speed up back-office tasks like answering performance questions in real time. It may also help with scenario analysis under various economic conditions.
- Investment banking and capital markets: It will generate stress-test scenarios for illiquid financial products to inform the proper compliance measures and reduce costs.
Generative AI has promising use cases over the next three years.
- Fraud support: Data created by generative AI can train algorithms to mitigate false positives and negatives that occur in fraud detection to improve accuracy and speed.
- Personalized offers: Generative AI can make individualized offers with images and natural language.
- Virtual assistants: The industry already uses virtual assistants and chatbots, and generative AI will help them respond to increasingly complex customer inquiries.
- Wealth planning: Generative AI will be able to simulate various client needs and economic scenarios so financial advisors can give situation-specific financial recommendations.
But generative AI won’t be effective in all areas of banking.
- Loan decisioning: Synthesized data generated by AI is hard to explain and difficult to audit, which may make it noncompliant in loan decisioning.
- Compliance: Because rules and regulations are constantly changing, generative AI likely won’t be able to simulate the proper regulatory environment. It may, however, eventually be used to draft statements of compliance or noncompliance.
- Trading: Generative AI likely won’t be used to train trading algorithms, as markets are often unpredictable.
Continue reading: To find out what banks should do to prepare for the rise of generative AI, check out our ChatGPT and Generative AI in Banking report.