The Taiwan Banker

The Taiwan Banker

How to control AI, and not be controlled by it

How

2023.04 The Taiwan Banker NO.160 / By Win Way Ngo and Chang Chung Hsia

How to control AI, and not be controlled by itBanker's Digest
Ever since the Dartmouth Workshop in 1956, artificial intelligence has received widespread attention and research. Over the past 60 years, there have been three progressive waves of AI, starting from simple logic and reasoning advancing onto visual recognition, natural language processing, machine learning, and deep learning. Recently, ChatGPT by OpenAI has made a breakthrough in natural language processing, and set a record of 100 million active users in less than 2 months after registration was opened at the end of 2022. Microsoft also announced a partnership with OpenAI at the beginning of this year, investing $10 billion to support its R&D, and at the same time connecting products such as its Bing search engine and Teams Premium. It recently added Copilot into its existing office software, opening up a new era of intelligent word processing. As a result, the public is paying great attention to the impact of AI, and people have been trying to use ChatGPT to improve their work efficiency, further exploring applications of generative AI in business, medical, financial and creative fields. Thus, 2023 has become known as the “year of generative AI.” AI can replace non-routine tasks As an article in the Economist magazine late last year noted, “Artificial intelligence is permeating business at last.” According to a McKinsey survey, 50% of the world’s companies have already tried to implement it. The Economist emphasized that it is a general-purpose technology that will impact overall economic and industrial development in the near future. Past general-purpose technologies such as steam engines, electricity, computers, and the internet have led to major transformations in industry and society. Everything is likely to become more efficient and precise thanks to AI, from automated production lines to intelligent customer service, medical diagnosis to financial analysis, and from smart farms to smart cities. The rise of generative AI has made it clear that AI can not only replace routine work through automation, but also some non-routine work – despite the need for a high degree of expertise, creativity, and social skills. The White House Council of Economic Advisers (CEA) published an article late last year exploring the impact of AI on the future workplace, finding that since it can perform non-routine work and tasks previous performed by humans, it may replace some existing jobs. At the same time, it will increase productivity and create new jobs, making it necessary for workers to learn new skills and change jobs. A European Parliament study also found that due to AI, the future labor force may undergo major shifts. Low-paying routine jobs are the most likely to be replaced, while high-skilled workers will be more productive due to collaboration with AI. The combination of these trends may lead to wage compression, higher unemployment, and a widening income distribution. Service customization for the financial industry From the perspective of the financial industry, AI can be applied in a number of areas, including credit assessment, risk control, portfolio management, and fraud detection, helping financial institutions to more accurately assess risk and value returns, thereby improving overall efficiency and profitability. Gartner identified generative AI as the top among the top three fintech trends in 2022, and forecasts that 20% of all consumer data training data in 2025 will be synthesized using generative AI, which can be applied to fraud detection, transaction prediction, and risk factor modeling through technologies such as generative adversarial networks (GAN), further enhancing the degree of customization in the financial industry. According to McKinsey, with the development of automation and AI, demand for manual work skills and basic cognitive skills will decrease significantly by 2025, while demand for social-emotional skills and advanced technical skills will increase significantly. The World Economic Forum’s Future of Work 2020 report also lists data scientists, AI experts, big data experts, digital marketing strategists, and process automation experts as the top five occupations for rapid growth, echoing McKinsey’s conclusion. For the financial industry, McKinsey argues that demand for bank tellers, accountants, and securities brokers will decrease significantly, while demand for technological experts and occupations requiring a high level of customer interaction will increase. Therefore, in the future, financial practitioners will need the ability to manage customer relationships, think creatively and collaborate across borders, and use AI and data analysis technologies. Information security and risk control will become critical In the past, creative work was considered irreplaceable, relying as it does on human creativity and imagination. With ChatGPT, this view is changing. Morgan Stanley, an international investment bank, was the first to announce in mid-March that it would use GPT-4 technology to create an in-house chatbot to help its large team of financial advisors access information quickly and produce insights on the latest trends, preserving more time to serve clients. However, some financial institutions have doubts about the security of ChatGPT. Bloomberg reports that a number of major banks have banned their employees from using it, including Bank of America, Citigroup, Deutsche Bank, Goldman Sachs, and Wells Fargo. One reason for this is data security: employees using ChatGPT may accidentally leak internal data, leading to trust issues and losses. Second is risk management: Although ChatGPT can improve efficiency, it may also generate problems such as ambiguous semantics or incorrect content, which may adversely affect risk assessment and decision making. McKinsey reports that while generative AIs like ChatGPT are currently in the early stages, this technology and other foundational models are gradually changing the AI landscape by providing more efficient assistance. Some of the early applications include customized marketing and social media content (including text, images, and video), text highlighting, assisting programmers in writing and reviewing code, drafting product brochures, analyzing customer feedback, and intelligent customer service. Since finance is a highly regulated industry, most banks are conservative about emerging technologies, but McKinsey argues that as technology and monitoring continue to advance, private ChatGPT services will emerge to ensure the security of internal data, and the accuracy of its feedback can be improved through continued model optimization. Better efficiency through AI The primary concern in this new wave of AI is not whether it will replace human work, but how to use it to improve work efficiency. Colleges and universities are already actively cultivating students’ understanding of AI, so financial practitioners in the workplace must keep up with future trends and continue to improve themselves. According to Deloitte’s survey Talent and Workforce Effects in the Era of AI, more than 50% of surveyed companies have already started training developers to create AI solutions, information personnel to deploy AI solutions, and employees to use AI in their work. This will enable employees to improve their overall work efficiency, and companies to thereby gain a competitive edge in the market. Looking back at history, structural manpower adjustments may occur in the process of transformation. New technologies replace old work, but they also create new types of work. Carriage drivers were replaced by drivers, and blueprints were replaced by CAD. You will not necessarily be eliminated by the technology itself, but by people who have mastered it. Fin & Tech Innovation Village will play a crucial role as a supply station for the future digital workforce. This year, several series of special courses are being offered in order to help students quickly adapt to the challenges of the AI era, including the Agility, Integration, and AI series. The Agility series aims to develop employees’ ability to quickly adapt to market changes, the Integration series teaches how to achieve effective cross-department and cross-business cooperation, and the AI series covers AI fundamentals, machine learning, and deep learning, helping financial practitioners master application of AI, and enhancing their digital literacy in order to better meet the challenges of the AI era. In addition, we should not overlook the importance of interpersonal communication and teamwork. Although AI has made significant progress in many areas, it is still cannot replace the advantages of human interpersonal communications and teamwork. While it cannot fully simulate human emotions, good interpersonal skills and teamwork will still give professionals an advantage in the workplace. Therefore, the Village also offers a course on consensus building for executives, the Cybersecurity and Information Security Executive Program (CISE), covering the core skills of leadership, decision-making, and strategic thinking to develop leaders with internal and external organizational coordination and strategic planning abilities. Through CISE, participants gain a deeper understanding of the importance of communication and teamwork within an organization, enabling them to meet the challenges of the digital era.