Polish Banks Begin Modeling How AI Could Reshape the Credit Market

Alior Bank and Erste Bank say they are not yet seeing any clear signals that advances in artificial intelligence are affecting customers' decisions to take out mortgage or cash loans. Even so, the financial institutions are starting to analyze what could happen to the credit market if automation begins to genuinely threaten the incomes of people working in occupations most exposed to being replaced by AI.

An article in Rzeczpospolita, citing data from NASK and the International Labour Organization, notes that more than 5 million jobs in Poland could be partially susceptible to automation using generative artificial intelligence. About 800,000 people currently work in occupations considered most exposed, chiefly in administration, accounting, customer service and data entry.

No panic for now

Angelina Szczepaniec, who oversees the rollout of AI solutions at Alior Bank, says plainly that the bank currently sees no clear signals that AI development is meaningfully affecting customer decisions on mortgage or cash loans. Szymon Kosnik, director responsible for retail customer risk at Erste Bank, offers a similar assessment, saying nothing suggests AI is increasing payment delays or reducing demand for credit.

We currently do not observe clear signals that AI development is significantly affecting customer decisions on mortgage or cash loans - Angelina Szczepaniec, Alior Bank

Today we see no clear impact of AI on the quality of our loan portfolio. No signals indicate that AI is increasing payment delays or reducing demand for credit - Szymon Kosnik, Erste Bank

Why banks are watching anyway

The mechanism that interests bank analysts is an indirect one. The greater the uncertainty about future employment and income, the less willing households are to take on long-term commitments, especially mortgages spread over twenty or thirty years. Someone working in an occupation considered at risk of automation might simply postpone a decision to buy a home on credit, regardless of whether they actually end up losing their job.

So far, though, sales data show no such effect. May 2026 was the third consecutive month in which mortgage sales in Poland exceeded 13 billion zloty. The main driver of demand remains concern over rising interest rates, inflation and further increases in property prices, not anxiety about job automation.

Credit scoring over worst case scenarios

Banks stress that they do not currently treat people working in occupations most exposed to automation, such as junior programmers, copywriters or call center staff, as higher risk customers. Creditworthiness assessment remains individualized, and AI's impact on the labor market is factored in at most indirectly, as one of many macroeconomic variables considered in scoring models.

That distinction matters, because it means banks are for now deliberately holding back from introducing additional filters based on a customer's occupation and its exposure to automation. Such a change would be legally and reputationally controversial, and could be seen as a form of occupational discrimination.

What comes next

The financial institutions say they will keep monitoring the situation and preparing various scenarios, but none of them forecasts a sudden collapse of the credit market anytime soon. They describe the likely path as evolutionary change rather than a sudden crisis, with the key factor being how quickly companies actually start replacing workers with AI systems, rather than announcements or technology pilots alone.

For people working in administrative, accounting or customer service roles, this means no direct consequences for now when applying for a loan, but it is also a sign that banks are taking the issue seriously enough to build internal analyses in case a scenario of mass automation starts to materialize faster than expected.

Sources: AI may make it harder for us to access credit. Banks are already monitoring the markets (rp.pl)