The financial landscape in 2025 has been dynamic, with trade tensions and policy shifts drawing intense market attention. In Hong Kong, the performance of local capital markets and the broader economy  have also been in focus. The rapid development of AI has been another major theme. We used three AI models with web-search capabilities, enabling them to access real-time information, to compile a preliminary ranking of the top financial news stories of 2025, both internationally and in Hong Kong. 

The following prompt was used for all three AI models:  "Please select the top 5 economic and financial news stories related to [Hong Kong/International] in 2025. You must use credible news websites and official news reports as information sources. For each selected news story, provide a clear ranking, detailed reasons for the ranking, and the source of the news. Ensure the information is accurate and up-to-date,” The list is presented in the table.

Duan Yang, a senior lecturer of the Department of Accountancy, Economics and Finance in HKBU,  noted that the results closely reflected key market developments, including the slow recovery of China’s real estate sector, weak consumer spending, the Hong Kong IPO surge, HSBC privatization, and Hong Kong government fiscal policies.

In terms of Hong Kong’s financial market, Duan observed that the city’s position as a financial hub had strengthened this year. While (Global Financial Centres Index) rankings had declined in previous years, strong activity in the overall financial market, particularly in the IPO sector, combined with substantial capital inflows, contributed to consolidating Hong Kong’s financial standing.

Duan also highlighted a notable trend in the Stock Connect program, with a majority of funds investing in Hong Kong equities now originating from mainland China, compared with a larger share from foreign investors in prior years, indicating a significant shift in market dynamics.

Professor Zhou Haonan, a finance professor at the University of Hong Kong, stated that one of the top global financial news stories in 2025 would likely be tariff policy developments. The tariff announcement in early April triggered significant market volatility, yet the US dollar unexpectedly weakened rather than strengthened, a clear indication that the immediate market impact of tariff news differs from traditional expectations.

Another major topic Zhou highlights is the development of stablecoins,  including legislative and regulatory progress, as well as official statements indicating a push for the digital yuan over stablecoins. Zhou also believes that new trends have emerged in the field of digital currency.

“For instance, major banks like Citibank are considering collaborating with stablecoin or cryptocurrency exchanges to assist clients in the conversion and circulation of stablecoins, cryptocurrencies, and fiat currencies such as the US dollar," He added.

In his view, the political process in France concerning its responsibility to service its sovereign debt,  and China’s new attempt to the internationalization of RMB, are also among the top financial news stories of the year.

For Hong Kong, Zhou pointed to several key financial developments in 2025, including progress on stablecoin‑related legislation and regulation, a rare episode in which short‑term interest rates fell below those in the United States under the linked exchange rate system, and a strong rebound in the IPO market, which he said “reflects the broader recovery in China’s economic activity.”

The questionnaire, which surveyed 30 college students, shows that 66.67% of respondents believe the ranking of financial news generated by ChatGPT is closest to their own judgment.

Ge Jiayi, one of the respondents, majoring in Public Relations and Advertising at Hong Kong Baptist University, focuses on the development of stablecoins in Hong Kong. She believes that the legislative and regulatory progress of stablecoins will have a profound impact on the digitalization of the financial market and cross-border payments, and therefore deserves close attention.

"With the licensing and regulation of stablecoins, using them is likely to become more convenient. For instance, activities like paying tuition fees or cross-border online shopping could become easier and cheaper in the future," Ge said. 

"Additionally, I think the Fed's interest rate decisions directly affect savings and loan rates, which in turn influence the pace of the global economy. The AI technology boom is also playing a significant role in shaping the economic landscape," she added.

 “The Sino-US trade war is the most crucial news because it may lead to rising import costs and prices, affecting business operations and family life,” Ruan Xinyan, a student from the Department of Journalism at Monash University, added.

AI Experiment in the Newsroom: Process, problems, and gains

For this special issue, our reporting team explored the use of different AI models and prompt words in various languages to generate the top 5 rankings of local and international news.

Using AI to rank news has been insightful. While AI can help organize and analyze information efficiently, results vary depending on the model, language, and training data. So, human judgment remains essential.

Prompt

At the beginning of using AI to generate rankings, our team used a relatively simple prompt: "Top 5 economic and financial news (with ranking) in 2025, and the reason?" However, Simon Wang, Lecturer and Innovation Officer at the Language Centre of HKBU, shared with us that “the result generated in this way is probably not very reliable”. 

So we refined the prompt: "Please select the top 5 economic and financial news stories related to [Hong Kong/International] in 2025. You must use credible news websites and official news reports as information sources. For each selected news story,  provide a clear ranking, detailed reasons for the ranking,  and the source of the news. Ensure the information is accurate and up-to-date."

We also created a Chinese version of the same prompt and submitted it to the three models on the same date to explore how language affects results.

Result

The ranking of the same topic of “Hong Kong’s Green Bond Issuance” has found significant differences across LLMs in both languages. In English prompts, ChatGPT listed “Digital Green Bonds” as fifth among Hong Kong stories, while “数位绿色债首发( Digital Green Bond First Issue)” rose to the fourth in Chinese prompts.

DeepSeek placed “成功发售第三批数码绿色债券 (Successfully issued the third batch of digital green bonds)” first on the Chinese prompts, while Perplexity completely waived this topic in both languages. These differences highlight that language input and semantic weighting vary across LLMs.

"The data used for training each model is different. If you ask it to search for sources online, the sources may also vary, which leads to different results," Fu Xiaoyi, professor of journalism at HKBU, explained.

Regarding the language-based discrepancies, Fu said, "This is because the training data behind Chinese and English, or the entire system and training data for training Chinese and English, may be different. Chinese uses Chinese and English uses English, and the data for the two may not be exactly the same."

“I'm not sure if (AI) can provide the best source of news,” Zhao said. "Since most mainstream financial media have paywalls, it is hard to say whether AI models obtain the content of these media and recommend relevant high-quality reports."  

Questionnaire

We also conducted a questionnaire among 30 college students from various disciplines, including Business, Journalism, and Data Science, to understand which language they prefer to use when writing prompts for AI.

We found that one-third of respondents prefer using English prompts, followed by those who have no preference between the two languages, then those who prefer using Chinese prompts. The smallest group was uncertain about their preference.