FICO has faith in AI as equitable force
Despite mounting concern over jobstealing artificial intelligence as machines take over Thailand’s lending industry, the analytics firm FICO sees AI as a force for good that could make lending not only safer and more efficient, but even more equitable.
“AI is used by data scientists to make lending more equitable by discovering ways to segment scorecards,” said Scott Zoldi, chief analytics officer of FICO. “Scorecards estimate whether a customer will exhibit positive or negative behaviours in the future based on credit history. Behaviours include the likelihood to repay a loan or whether they will miss a loan payment or become bankrupt.”
According to International Data Corporation, Thailand has the second-highest percentage of AI adoption for organisations in Asean, while spending on AI in Asia-Pacific (excluding Japan) is expected to reach US$4.6 billion
(151 billion baht) by 2021.
AI is being used by Thai banks to automate lending decisions, debt management and client communications (e.g., with chatbots).
The technology gives banks more tools to understand lending risk. Traditional methods place customers in rigid categories such as new and existing, while the AI models are more customisable and adaptive.
Mr Zoldi compared the old ways to animated cartoons, while the new AI models are like richly detailed 3D-rendered characters. Think Homer Simpson versus Thanos from the new Avengers movie.
While these new automated methods will replace some human workers, Mr Zoldi said new opportunities will be created for IT professionals who are proficient in AI.
A recent FICO forum held in Bangkok revealed that 40% of banks said the lack of talent with AI proficiency was the greatest obstacle to adopting new AI and machine learning.
“AI can actually help free employees from manual data collation work and allow employees to have more bandwidth to concentrate on strategic tasks,” Mr Zoldi said. “For example, when a company deals with debt collection, they can automate a portion of their collection efforts and reduce the cost and time taken to collect.”
To ensure a better AI future, FICO is developing what it calls “explainable AI”. Much of the AI developed is considered to be a black box, where even its creators don’t know why the AI makes certain decisions.
Explainable AI strives to use algorithms to help humans understand the decisionmaking process of machine learning models. This will ideally let humans drive the AI to make more ethical decisions, Mr Zoldi said.
Despite the confluence of government regulations and well-meaning developers, not every stakeholder is convinced that the proliferation of AI will be a good thing for mankind. Tesla founder Elon Musk regularly makes foreboding statements about AI’s potentially fearsome ramifications. In March he compared the danger of AI to that of nuclear weapons.
But Mr Zoldi, for one, is not afraid. “It’s important to take cautions against the risks associated with the adoption of AI technology,” he said. “But with the aid of regulations, all will be held to responsible use of AI.”