Artificial Intelligence and the Future of Hong Kong

Artificial intelligence (AI) has advanced dramatically in recent years. In particular, deep learning and generative AI models have gained widespread popularity at both organizational and individual levels, with a range of innovative and viable applications being developed.


1.    The Rise of AI

Artificial intelligence (AI) has advanced dramatically in recent years. In particular, deep learning and generative AI models have gained widespread popularity at both organizational and individual levels, with a range of innovative and viable applications being developed. Deep learning techniques focused on classification and prediction have substantially improved the performance of various fundamental AI tasks such as speech, image, and facial recognition, resulting in many real-world applications including identity verification, self-driving cars, medical diagnosis, and supply chain management. Meanwhile, generative AI has widened access to sophisticated content creation tools for texts, images, and videos, particularly the easy-to-use ChatGPT released by OpenAI in November 2022 and similar chatbots based on large language models. These developments have fuelled worldwide interest in AI across government, industry, academia, and the general public.

While AI brings promising applications to society, it also presents significant challenges. For example, many people fear AI-driven automation will disrupt some industries, eliminating certain job categories and potentially causing mass unemployment. Organizations and people must therefore adapt and prepare for AI’s rapid growth. Another area of concern is AI safety and  responsibility. The legal consequences of AI’s actions  have not been clearly defined in many situations involving fast-growing applications. Criminals are also using AI to conduct sophisticated scams and other illegal activities. This article explores three key dimensions linked to AI, namely Hong Kong’s competitiveness, the labour market, and safety considerations. Following these discussions, we make several policy recommendations related to education, funding mechanisms, and regulatory frameworks.

2.    AI and Hong Kong’s Competitiveness

AI has seen its fair share of ups and downs in past decades. The rise of generative AI in 2022 symbolizes the industry’s current boom, but its foundation lies in deep learning and large language models that have been gradually gaining momentum since the 2010s. AI’s renaissance today has yielded breakthroughs across virtually every domain, from scientific research and medicine to engineering, business, and the arts. 

To maintain Hong Kong’s competitiveness globally, it is important for all sectors to embrace AI. While AI’s potential is particularly evident in areas such as financial technology, e-commerce, marketing, logistics, healthcare, and smart city solutions, its applications extend far beyond these industries. The city must think about where and how it can strategically use AI to boost productivity and efficiency across society.  By leveraging AI technologies, Hong Kong can improve its business environment, attract more investment, and foster economic growth. Without such planning, we will easily lag behind our peers and competitors.

Many people have asked: Can Hong Kong become a leading AI hub? Hong Kong offers a number of  advantages, including its free-market economy, business-friendly environment, access to the Greater Bay Area and the China market, and a diverse talent pool. But Hong Kong also faces fierce competition  from regional rivals such as Shenzhen, Taiwan, and Singapore. Or worse, Hong Kong is caught in the technology war between China and the United States. For example, OpenAI has suspended services to users in Hong Kong, while Nvidia has banned export of powerful computer chips to Hong Kong and mainland China. These moves severely hamper Hong Kong’s potential role  as a regional leader in AI.

To address these challenges, Hong Kong must forge a distinctive strategy and consider alternatives. One solution is to develop our fundamental research capabilities. While some prominent applications like AlphaGo and ChatGPT were developed by large organizations with strong financial backing, startups can still succeed and achieve breakthroughs in this field. One notable example is SenseTime, a leading AI company founded in Hong Kong that focuses on image recognition.[1] Building our own strengths in both basic research and real-world applications is crucial  for keeping Hong Kong ahead in the global competition.

3.    AI and the Labour Market

AI has the ability to fully or partially automate many  tasks in the job market. The International Monetary Fund estimated that AI will impact nearly 40% of global employment, with that figure reaching up to 60% in advanced economies like Hong Kong.[2] Job categories at high risk of being disrupted or even replaced by AI include data entry and administrative jobs, customer service, manufacturing and assembly lines, retail checkouts, analytical roles, graphic design, translation, and commercial photography.[3] Others have suggested that vehicle drivers and jobs that involve content creation, such as social media marketers, computer programmers, and technical writers, are also particularly vulnerable.[4]

While technology has upended labour markets before, what’s distinct about this AI wave is its likely impact on high-skilled, high-income jobs, compared to earlier waves of industrial automation that mostly affected lower-wage occupations.[5] Optimistically speaking, AI will substantially enhance the productivity of high-income workers, allowing these individuals to direct their time to more value-adding work that could only be done by humans. However, a more sobering perspective suggests that many high-income professionals will be displaced from their jobs. Without appropriate training, these individuals may find it even harder than low-income workers to adapt to the AI era, as AI may now be able to perform their once-valuable skills at a much lower cost.

At the same time, AI can create entire job categories, like how the Internet revolution brought about new ones like blogging and live-streaming. Examples of novel professions  include prompt engineers, AI experience designers, and AI trainers. In any case, AI will profoundly impact global and local labour markets, replacing and creating many jobs in a complex interplay.

4.    Responsibility and Safety of AI

AI can automate many tasks currently performed by humans, and is expected to take charge of even more functions without human intervention. Examples include verifying identities, managing an investment fund, driving a car, or even making medical diagnoses. Any missteps could result in disastrous consequences, from investment losses to traffic accidents involving casualties.[6] AI systems, like other information technologies, are subject to hacking, viruses, and software bugs. AI may also suffer from bias and discrimination inherited from the data it was trained on. These issues, which could lead to unintended and costly results, make clear that AI poses safety issues to society. Moreover, bad actors may take advantage of AI; for example, a company was scammed HK$200 million through a video meeting produced by deepfake technologies.[7] Governments must therefore educate the public about the power and risks of AI, and devise regulations and guidelines on the responsibilities and ethical considerations of AI usage.

Since AI technologies and applications are advancing rapidly, laws and regulations are often playing catch-up, like in the autonomous driving industry. Regulators must strike a balance between providing a legal basis for the latest AI technologies and clearly defining their responsibilities and boundaries.

5.    Policy Recommendations

To prepare for the AI era, Hong Kong could consider the following recommendations:

  1. AI Education: AI literacy, competencies, and its responsible uses should be woven into education at all levels. Citizens should understand the benefits, risks, and limitations of AI from a young age. Primary and secondary schools can incorporate AI into formal curriculums. Authorities and educators can set up a council dedicated to boosting AI literacy through public seminars, workshops, television and online videos, and other activities. These initiatives will allow the general public to use AI more effectively and responsibly and better prepare the workforce for the changing dynamics in the labour market.
  2. Help Organizations Get On Board with AI: Many organizations, especially small and medium enterprises and non-government organizations, are eager to apply AI to improve their productivity, but do not have the resources or expertise to do so. The government can provide assistance, such as training workshops and funding focused on AI, to help these organizations.
  3. Support Frontier AI Research: With a wide range of applications, AI is critical to Hong Kong’s future economic success. To maintain global competitiveness, the government must support cutting-edge AI research and development across universities and startups. Increased funding for both fundamental research and practical applications will not only accelerate Hong Kong’s AI capabilities but also create high-value job opportunities.
  4. Attract Non-local Investments on AI: Hong Kong should attract non-local companies, including both mainland China and overseas companies, to develop and apply advanced AI technologies in the city. Various incentives and subsidies can be provided to non-local companies to establish offices and businesses in Hong Kong. At the same time, authorities should facilitate non-local investments into homegrown AI companies to bring in expertise, resources and collaboration opportunities, fostering AI’s growth in the city.
  5. Regulations and Guidelines on AI: Given the ethical, legal, and safety challenges associated with AI, the government should continuously review and update existing AI and data-related laws in Hong Kong. New regulations or amendments should be introduced as needed in response to the rapid advancements in AI. These laws should also consider frameworks from other advanced economies and mainland China. More industry-specific regulations and guidelines, such as those published by the Hong Kong Monetary Authority[8] on AI usage, can help different sectors embrace AI responsibly.

[1] https://www.sensetime.com/

[2] https://www.imf.org/en/Blogs/Articles/2024/01/14/ai-will-transform-the-global-economy-lets-make-sure-it-benefits-humanity

[3] https://www.forbes.com/sites/bernardmarr/2024/06/17/what-jobs-will-ai-replace-first/

[4] https://www.techtarget.com/whatis/feature/Will-AI-replace-jobs-9-job-types-that-might-be-affected

[5] https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-AI-the-next-productivity-frontier

[6] https://www.bbc.com/news/technology-67133409

[7] https://www.scmp.com/news/hong-kong/law-and-crime/article/3250851/everyone-looked-real-multinational-firms-hong-kong-office-loses-hk200-million-after-scammers-stage

[8] https://www.hkma.gov.hk/media/eng/doc/key-information/guidelines-and-circular/2024/20240819e1.pdf

Translation

人工智能與香港前景


 

引言


 

近年來,人工智能取得了飛躍的進展,尤其是深度學習和生成式人工智能模型,在機構和個人層面同樣大受歡迎,並開發出一系列創新且可行的應用。聚焦於分類和預測的深度學習技術,大大提高了各種基本人工智能任務的性能,如語音、圖像和容貌辨識,從而在核實身分、自動駕駛汽車、醫學診斷和供應鏈管理等多方面獲得實際應用。與此同時,生成式人工智能面世後,創作精緻文本、圖像和視像等內容的工具用途得以擴充,特別是在2022年11月,由OpenAI發布易於使用的ChatGPT和基於大型語言模型的同類聊天機械人。這些發展激發了各國政府、產業、學術界和公眾對人工智能的廣泛興趣。

雖然趨勢揭示人工智能在社會上有望大派用場,但也帶來了重大挑戰。舉個例子,許多人擔心人工智能驅動的自動化操作會顛覆部分行業,令某些工種消失,並可能導致大規模失業。因此,機構和個人必須加以適應,為人工智能的快速增長做好準備。另一個關注點是安全性和責任問題。在許多涉及人工智能用途激增的情況下,人工智能行為的法律後果尚未明文規定。犯罪分子也在利用這種新技術精心設計出種種騙局,並進行其他非法活動。

本文探討與人工智能相關的三大層面,即香港的競爭力、勞動力市場和安全考量;並就教育、資助機制和監管框架提出幾項政策建議。

 

人工智能與香港的競爭力


 

人工智能在過去幾十年經歷了不少起伏。2022年,生成式人工智能崛起,象徵這個行業當前的熱潮,但其基礎實繫於自2010年代以來持續進展的深度學習和大型語言模型。今時今日,從科學研究、醫學到工程、商業和藝術,人工智能復興幾乎在每個領域都取得了突破。

要保持香港在全球的競爭力,各行各業都不能不擁抱人工智能。雖然其潛力在金融科技、電子商務、市場營銷、物流、醫療保健和智能城市解決方案等範疇尤為明顯,但其用途遠不止於此。香港必須考慮在哪些行業以及如何策略性地加以運用,從而提升整體生產力和效率。借助人工智能技術,足可改善營商環境,吸引更多投資,並促進經濟增長。若欠缺此等規劃,難免落後於各地城市和競爭對手。

不少人問:香港能夠成為領先的人工智能樞紐嗎?特區優勢眾多,包括自由市場經濟、友善營商環境、作為大灣區和全國市場的窗口,以及多元化的人才庫;但與此同時,亦面對來自深圳、台灣、新加坡等區內競爭對手的激烈競爭。更有甚者,香港被捲入中美之間的科技戰,例如OpenAI已暫停對香港用戶的服務,而Nvidia則禁止向香港和中國內地出口高端電腦晶片。這些舉措無疑窒礙香港在區內人工智能技術領域發揮其領導地位。

為了化解上述挑戰,香港必須制定獨特的策略,並審視替代方案。方案之一是培育本地的基礎研究能力。雖然一些應用(如AlphaGo和ChatGPT等出色的系統)由實力雄厚的大公司開發,但初創企業仍然可以在這一環取得成功和突破。顯例之一是在香港創辦的商湯科技,這家龍頭人工智能公司以圖像識別為核心技術。[1]要在全球競爭中佔先,香港務須在基礎研究和實際應用方面厚積實力。

 

人工智能與勞動力市場


 

在就業市場中,不少職務都可以利用人工智能完全或部分自動化。國際貨幣基金組織估計,人工智能的影響將及於全球職場近乎40%,在類似香港的發達經濟體中,這一比例更可高達60%。[2]被人工智能衝擊甚至取代的高風險工種,計有數據輸入和行政工作、客戶服務、製造和組裝生產線、零售結帳、分析崗位、平面設計、翻譯和商業攝影。[3]亦有意見認為,易受池魚之殃的工作包括車輛駕駛員和涉及內容創作的職位,如社交媒體營銷員、電腦程式編製員和技術寫作員。[4]

雖然科技顛覆勞動力市場早有先例,但這次人工智能浪潮與別不同之處,在於它或對高技能、高收入工作造成影響,反觀早期一波又一波工業自動化浪潮,受影響的主要屬低收入職位。[5]樂觀而言,人工智能將大大拉高高收入者的生產力,他們就能夠將時間花在只有人類才可勝任的較高增值工作之上。然而,換一個發人深省的角度來看,不少高收入專業人士將被取代。在缺乏適當培訓的大前提下,他們可能會發現比低收入的就業人口更難適應人工智能時代,因為人工智能現在或能以極低的代價,就可掌握他們昔日的寶貴技能。

話說回來,人工智能亦會創出整套全新的工種,好比互聯網革命產生了博客和串流直播等新工種一樣。新興專業包括提示工程師、人工智能體驗設計師和人工智能培訓導師。無論如何,人工智能將深遠影響國際和本地的勞動力市場,在複雜的相互作用環境中,取代和創造林林總總的工作。

 

人工智能的責任與安全


 

事實上,目前多種由人類執行的任務,都可藉人工智能加以自動化,預計將來在無人工介入的情況下,處理更多職能,例如核實身分、管理投資基金、駕駛汽車,甚至作出醫療診斷。從投資損失到釀成傷亡的交通事故,任何失誤都可能導致災難性後果。[6]與其他資訊科技一樣,人工智能系統也會遭受黑客攻擊、病毒和軟件程式錯誤的問題。基於所接受訓練的數據中存在偏見和歧視,人工智能或會備受影響。凡此種種可能導致意想不到且代價高昂的結果,可見人工智能對社會構成安全風險。再者,不法分子或會利用人工智能作姦犯科;曾有一家公司誤信由深偽技術製作的視像會議,最後損失2億港元。[7]政府責無旁貸,應教育公眾有關人工智能的威力和風險,並制定人工智能使用責任和倫理考量的規例和指引。

人工智能技術和應用正發展得如火如荼,法律和規例往往處於從後追趕狀態,從自動駕駛行業的情況可見一斑。為了給最新的人工智能技術提供法律依據與明確界定法律責任和範圍,監管機構必須在兩者之間取得平衡。

 

政策建議


 

為了做好準備,迎接人工智能時代,香港不妨考慮下列措施:

  1. 普及人工智能教育:人工智能素養、能力及負責任用途,都應該一一融入各個教育階段。讓公民從小學習人工智能的效益、風險和局限性;小學和中學應將人工智能納入正規課程。當局和教育工作者可以設立一個專門的委員會,通過公眾研討會、工作坊、電視和網上視像以及其他活動,以便提升人工智能素養和教育。這些舉措將有助市民大眾更有效和負責任地使用人工智能,並協助工作人口為勞動市場的動態變化做好準備。

  2. 扶助機構採納人工智能:很多機構(尤其是中小企業和非政府組織)都渴望應用人工智能來增強生產力,卻礙於缺乏資源或專業知識而無能為力。政府可透過人工智能為題的培訓工作坊和資助等方式,予以協助。

  3. 驅動前沿人工智能研究:人工智能用途廣泛,對香港未來的經濟成就舉足輕重。為了保持全球競爭力,有賴政府支持大學和初創企業的尖端人工智能研究和開發。增加對基礎研究和實際應用的資金投入,不僅會催化香港的人工智能發展,還會創造高價值的就業機會。

  4. 就人工智能招商引資:香港應設法招徠中國內地和海外公司,在本地開發和應用先進的人工智能技術。當局宜提供各項獎勵措施和補貼,以吸引此等公司來港設立辦事處和經營業務。同時應為本地人工智能公司引入境外投資,打造專門知識、資源和合作機會,引領人工智能的發展更上層樓。

  5. 修訂人工智能的規例和指引:鑑於人工智能在倫理、法律、安全方面的挑戰,特區政府應持續檢視和更新現有與人工智能和數據相關的法律條文。人工智能技術發展一日千里,理應適時制定新規例或修訂現行法例,過程中應參考其他先進經濟體和中國內地的框架。其他各項專屬行業的規例和指引,例如香港金融管理局[8]發布的人工智能使用指導原則,亦有助各行各業負責任地採納人工智能。


 

[1] https://www.sensetime.com/

[2] https://www.imf.org/en/Blogs/Articles/2024/01/14/ai-will-transform-the-global-economy-lets-make-sure-it-benefits-humanity

[3] https://www.forbes.com/sites/bernardmarr/2024/06/17/what-jobs-will-ai-replace-first/

[4] https://www.techtarget.com/whatis/feature/Will-AI-replace-jobs-9-job-types-that-might-be-affected

[5] https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-AI-the-next-productivity-frontier

[6] https://www.bbc.com/news/technology-67133409

[7] https://www.scmp.com/news/hong-kong/law-and-crime/article/3250851/everyone-looked-real-multinational-firms-hong-kong-office-loses-hk200-million-after-scammers-stage

[8] https://www.hkma.gov.hk/media/chi/doc/key-information/guidelines-and-circular/2024/20240819c1.pdf