Before a company’s loan request reaches a room full of bankers, a computer may have already sized it up.

Complete the form to unlock this article and enjoy unlimited free access to all PYMNTS content — no additional logins required.

Artificial intelligence tools now score borrowers, sort through financial statements and write first drafts of loan approval memos. In many cases, nobody at the bank formally signed off on that change. That could become a big problem when someone asks how a loan decision was made.

That’s the warning from Outside GC, a firm that provides outside legal counsel to companies, in an article by Partner Kristin Kreuder. AI often slips into the lending process through the back door. A bank might buy a tool to organize a borrower’s financial data. Over time, the tool starts flagging problems with loan terms. Then it helps write the approval memo. Before long, the bank’s risk ratings start moving with whatever the tool says. Nobody updated the bank’s lending rules to account for it.

The drift creates real legal risk. Banks must be able to explain loan decisions to regulators, borrowers and other lenders, the article said. If a loan goes bad, someone will want to know what the AI reviewed, what it missed and where a person stepped in. A memo that reads like a computer summary may not hold up under that kind of scrutiny.

The rules aren’t simple either. Federal banking regulators updated their guidance on how banks manage risk models in April, according to the article.

Source link