香港人工智能格局——跨行业调查

Full Report Partners of the Study WOW Impact Group Hong Kong FinancialIndustry Association Hong Kong DigitalFinancial Association HK01 Association of Cloudand MobileComputingProfessionals e:contact@wowimpact.com: +852 9603 3026 Study Objective This surveyaims to examine stakeholder perspectivesin Hong Kong on the adoption of Generative AI andAgentic AI, responsible AI practices, and the policylandscape, with a focus on collaboration amonggovernment, industry, academia, research, application,and investment sectors. The key objective is to identify gaps and proposesolutions to enhance Hong Kong’s innovationecosystem as a global financial and technological hub. The survey was conducted in July 2025, and a total of1343 responses have been received. Respondent Profile Bridging Hong Kong’s AI Gaps: A Cross-sectorAnalysis of Infrastructure, Policy, and Talent forGlobal Competitiveness Hong Kong stands at a critical juncture in its pursuit ofbecoming a leading innovation hub, particularly in theburgeoning field of Artificial Intelligence (AI). This reportidentifies significant challenges that threaten the city’s globalcompetitiveness in AI development and adoption.Keyfindings reveal urgent gaps in AI infrastructure and policy,pervasive talent shortages, and a pressing need for robust,adaptive governance frameworks, especially for autonomousAI systems. The AI innovation ecosystem is further constrained by apronouncedtalent shortage, exacerbated by internalworkforce unpreparedness and resistance to AI adoption.This indicates a dual challenge of attracting new expertiseand upskilling existing human capital. The imperative forresponsible AI governance is underscored by fragmentedawareness of AI risks, low organizational preparedness for AIpolicies, and a dangerous lag in updating governanceframeworks, particularly concerning the ethical implicationsof autonomous systems. A substantial portion of surveyed organizations perceiveshigh infrastructure costs and a lack of clear policy guidanceas major impediments. This is compounded by limited andextensive adoption of specialized AI infrastructure,suggesting a foundational bottleneck. Furthermore, while thetransformative potential of Generative AI (GenAI) and AgenticAI is recognized across public services, BFSI, and healthcare,cross-sector collaboration remains underdeveloped,hindered primarily by policy uncertainties and data sharingrestrictions. Despite the public sector Smart Cities initiatives leading AIadoption, practical deployment readiness remains low, withsignificant policy and ethical hurdles demanding immediateattention. To secure its position as a global AI leader, HongKong must prioritize strategic investments in accessible AIinfrastructure, enact clear and adaptive regulatoryframeworks, cultivate a deep and broad AI talent pool, andfoster genuine cross-sector collaboration grounded in trustand shared vision. AWARENESS ANDUNDERSTANDING OFGEN AI AND AGENTIC AI 01 There is a moderate level of familiarity with AI technologies across various keysectors in Hong Kong. Generative AI is the most widely recognized, while Agentic AIremains less understood. Professionals in technology-focused sectors and those inlarger organizations demonstrate a higher degree of familiarity. The prevailing view isthat these technologies will bring incremental efficiency gains rather thanrevolutionary disruption to most industries. Familiarity with AI technologies There is a medium level offamiliarity across the board.Generative AIis the mostwell-known technology, with44% of respondentsindicating they are familiar orvery familiar with it. Agentic AIis the leastunderstood, with only 30% ofrespondents expressingfamiliarity. Predictive analytics andcomputer vision fall inbetween, with familiaritylevels of 33% and 34%,respectively. Familiarity with AI technologies–Sector Affiliation Individuals who are more directly involved in the practical application, governance, and research of AI have a stronger understanding of thesetechnologies. Familiarity with AI technologies Unsurprisingly, theTechnology & ArtificialIntelligence sectordemonstrates the highestlevel of knowledge across allAI categories. In contrast, theRetail & e-commerce sectorshows the lowest familiarity. The data also reveals thatfamiliarity with AI tends toincrease with the size of theorganization, with largercompanies reporting higherlevels of understanding,especially for Generative AI. Impact ofGenAIand Agentic AI The predominant view, heldby 50% of all respondents, isthat these technologies willlead to incrementalefficiency gains. A smaller but notablepercentage, especially in theInvestment and Researchfields, anticipate arevolutionary disruption. The belief in incrementalefficiency gains is strongacross most industries,particularly in Smart Cities &Public Services, and Finance &Insurance, as well as in thelarger organizations. The Technology & AI andFinance & Insurance sectorshave the highest e