UK banks: what AI regulation could look like
UK banks have been warned by financial regulators not to use artificial intelligence systems to approve loan applications unless they can prove their algorithm is not biased against minorities, according to the Financial Times .
At this time, the warning does not take the form of an official ruling, and it leaves wide latitude for banks to continue using AI or machine learning systems as long as lenders ensure that the data used to feed the algorithms and the result does not discriminate against people who already have trouble borrowing. This post against algorithmic bias may be a preview of what the next white paper on AI regulation might look like.
Financial institutions around the world are using artificial intelligence models and machine learning to decide whether or not to grant loan applications based on the data they can collect, which in many cases includes zip codes, default rates in the region, benefits, salaries, etc.
The debate about using this data and allowing an algorithm to provide a final decision without human supervision is that the data fed into the algorithm may already be biased and it may skew a decision towards a discriminatory outcome. . Some of the information may be based on historical data rather than personalized data, which may unfairly affect a candidate. For example, living in a zip code where people are likely to default on a loan may affect your score for getting your application approved, even though your personal circumstances may be different than average. Other types of demographic information may have a similar effect on your application, unless you can demonstrate that you do not belong to this category – but without human supervision, such demonstration is no longer possible.
One way to improve the AI system is to include more data points to “personalize” the decision as much as possible, trying to eliminate or minimize the risk of biased decisions. Perhaps the best example of this is Alibaba’s Ant Group. Ant’s AI system automatically sets credit limits, interest rates, and even makes decisions based on the history of using Alibaba’s services. This means that before making a decision, Ant analyzes up to 3,000 data points for each consumer, including phone bills, consumer behavior and demographics. As a result, Ant’s decisions are unlikely to be biased – but if a company in Europe or the US tried to collect similar data, they would likely face privacy issues, as consumers would likely not feel comfortable disclosing this data.
Banks argue that it is the human factor, not the AI system, that is most likely to be subjective and provide unfair results. Both the AI system and the human factor can have flaws, but in coordination, they can also bring the best result, yet. In October, the Bank of England and the Financial Conduct Authority discussed an ethical framework and training around AI, including some human oversight and a requirement that banks could explain the decision taken by automated systems.
Read more: UK seeks its place to shape global standards in artificial intelligence
This is exactly what regulators are asking banks to do now: keep improving AI systems, as they have clear benefits for consumers and banks, but be careful about the datasets used and the results produced.
This warning may be the first step in AI regulation in the UK, which uses a soft approach with more guidance and less regulation. However, the government is expected to publish a white paper on AI regulation in early 2022. The white paper, which could eventually become law, will likely provide more information on how to reduce “bias” in AI . Algorithmic biases are probably the biggest concern regarding AI regulation among regulators around the world. The European Commission has proposed legislation in 2021 aimed at curbing this problem, and the US Federal Trade Commission also announced last year that it may take steps to reduce discriminatory outcomes when companies use AI. .
Read more: FTC considers new AI regulations to protect consumers
Register here for daily updates on the legal, policy and regulatory issues shaping the future of the connected economy.