離開
進入《TAIPEI TIMES》 FSC asks banks to monitor use of artificial intelligence
Financial Supervisory Commission Chairman Thomas Huang speaks at a meeting of the legislature’s Finance Committee in Taipei yesterday. Photo: CNA
By Kao Shih-ching / Staff reporter
The Financial Supervisory Commission （FSC） yesterday asked local banks and financial services providers to closely monitor their use of artificial intelligence （AI） tools such as ChatGPT, to protect data security.
The FSC told the Bankers’ Association （銀行公會） to investigate how many banks are using AI tools, how they use them and whether they have set risk controls, and to complete the review in two months.
Meanwhile, the Securities and Futures Bureau said it would investigate whether local securities companies have incorporated AI tools into their services, bureau Deputy Director Kao Ching-ping （高晶萍） told the Taipei Times by telephone yesterday.
Lawmakers on Thursday questioned whether the FSC was aware of the potential hazards involved in the use of large-language models, commonly referred to as AI.
No foreign financial regulator has taken a clear stance on the issue, although the Hong Kong Monetary Authority said the issue would eventually be discussed by international agencies, such as the Bank for International Settlements, FSC Chairman Thomas Huang （黃天牧） told legislators.
The FSC would research the risks to consumers first and decide how to set a framework, Huang said.
Banks seeking to use ChatGPT might risk data leaks, as they would be using customer information to interact with the software, which could incorporate that information into its own database, Banking Bureau Director-General Sherri Chuang （莊琇媛） said.
Even if banks use a special version of a large-language model for business, they should be aware of the data security risks and take precautions, Chuang said.
The Securities and Futures Bureau is unaware of any local securities company using ChatGPT as part of their equity research or customer service, but it would seek to confirm that, Kao said.
“There is no regulation banning securities companies from using new technology tools, but if the companies want to utilize the bot [ChatGPT], they must prevent data insecurity and be responsible for any financial advice they give to their clients, whether the advice based on their employees’ judgements or that of bots,” Kao said.
With ChatGPT gaining popularity worldwide, experts have been discussing whether it would replace financial services provided by humans, as the tool could analyze vast volumes of data quickly, and detect patterns and trends difficult for a person to spot.
However, when asked for advice on investing in local companies, the chatbot gave a neutral and conservative response.
It said that it cannot provide investment advice, which requires a thorough understanding of the investor’s financial situation, risk tolerance, investment goals and other factors.
Although some shares have performed well in the past, that performance is not a guarantee of future results, it added.
新聞來源：TAIPEI TIMES
