Thailand is rapidly developing artificial intelligence (AI) for fraud detection, according to SAS Institute, a global leader in the data and AI field.
The country's financial sector is expected to see risk management and fraud management collaborating to tackle new financial crime in an end-to-end prevention process, the company notes.
Thailand has strong regulatory frameworks and collaboration among banks. This aligns with successful models in Australia, Europe and Hong Kong in combating financial crime, said Ian Holmes, SAS director and global lead for enterprise fraud solutions.
Mr Holmes added that Thailand is embracing technology to combat financial crime, including real-time payment protection to safeguard against fraud in digital transactions.
The opening of new accounts is now subject to a stricter checking process to prevent scams.
Fraud and financial crime are not only banking issues but also involve sectors such as telecommunications, as fraudsters often use compromised phone numbers for social engineering attacks.
A growing threat is authorised push payment fraud, where customers are tricked into sending money to fraudsters. This is particularly challenging for banks as customers unknowingly become complicit in the transactions, Mr Holmes said.
He added that the wider adoption of AI is transforming the fraud detection and anti-money laundering processes. The use of Thailand's digital ID will also speed up identity verification, requiring AI to ensure secure and fast processing.
Mr Holmes said generative AI, including large language models (LLMs), is being tested for automating compliance tasks such as writing regulatory reports and case notes, which can reduce the manual labour involved in reporting.
Successful AI integration requires alignment with people and processes. Proper governance of AI and understanding of its data inputs are key to achieving return on investment, he added.
David Asermely, SAS's global lead of model risk management and AI governance, said banks and businesses are preparing for broader AI usage, specifically in the context of using LLMs for different tasks.
One example is the use of two models: a primary LLM to answer questions, and a separate "judge" model to grade the answers.
Banks are using AI to validate and test their own AI systems. After the judge model gives a grade, human experts review the results to ensure proper safeguards and accuracy are in place. This combination of AI and human expertise helps build better guardrails for the AI systems.
"There is estimated investment in AI projects, of which AI governance accounts for 5-7% of total spending," said Mr Asermely.
Nutapone Apiluktoyanunt, managing director of SAS Thailand, said the company foresees opportunities from the investment of financial and telecom companies in fraud detection and risk management solutions, in particular to tackle online scams, and in the future the development of virtual bank businesses.
SAS focuses on bank solutions in fraud and financial crime compliance, risk management and customer intelligence, he added.
The company has invested US$1 billion in the past three years to develop AI-powered solutions, he said.