Innovation facilitators, such as regulatory sandboxes and innovation hubs, have been widely adopted in Asia to foster innovation in the financial sector and in other industries. More recently, these facilitators have been used to support artificial intelligence (AI) innovation. This reflects the growing interest in harnessing the potential of AI technologies and to mitigate their potential risks. This chapter maps the evolution of these facilitators across Asia, discusses their benefits and the challenges they pose, and outlines common trends. It also discusses the key role innovation facilitators can play in capital markets. The chapter closes with policy considerations on balancing AI innovation with market stability and ensure their safe and beneficial use for all market participants.
Asia Capital Markets Report 2025
Abstract
AI-related activity has grown strongly across Asia and is expected to continue growing in the medium term. Levels of M&A activity related directly or indirectly to AI (e.g. data centres) and venture capital investments in AI-related start-ups are high in the region. Data centre capacity is expanding rapidly, with demand for AI computing power projected to grow ten-fold between 2023 and 2030.
Innovation facilitators are playing an increasing role in promoting experimentation with AI innovations in the region. One jurisdiction (Hong Kong, China) has introduced an innovation facilitator focused on promoting AI innovation in the financial sector, and in the rest of Asia another 11 facilitators have either demonstrated use cases of AI in finance or explicitly include AI within their scope. Six more facilitators are focused on AI innovations beyond the financial sector.
While Asian economies have taken different approaches to implementing innovation facilitators, some common trends are emerging. These include a relatively limited number of AI use cases across facilitators, a general preference for introducing sector- or product-specific facilitators, and a larger proportion of sandboxes compared to innovation hubs. On the other hand, there is significant divergence in the design of facilitators, in the level of public disclosure, and in the additional services or support provided to participating entities.
Innovation facilitators can play an important role in supporting a responsible and safe integration of AI innovations in financial markets, consistent with the OECD AI Principles. They foster closer collaboration between market participants and authorities, helping to address regulatory barriers or gaps and sending a positive signal about the commitment to responsible innovation.
However, they can also pose challenges, in some cases creating an uneven playing field by giving some companies a competitive advantage. Innovation facilitators can also increase risks of regulatory fragmentation and market distortions, while the resourcing requirements can impose opportunity costs. Lastly, the potential lack of AI expertise and knowledge can hinder regulators’ understanding of innovations and reduce the effectiveness of a facilitator.
Well-designed innovation facilitators, as well as international co-operation to promote the adoption of responsible and safe AI tools, can contribute to capital market growth. They can strengthen investor confidence, by improving the communication between authorities and the markets and encouraging safe testing of AI innovations. This can enhance access to finance for AI start-ups and foster the innovation ecosystem, ultimately promoting more dynamic and globally competitive markets.
The pace of innovation in the field of AI has accelerated in recent years, and Asian economies have been at the centre of this trend. More recently, GenAI applications have grown in importance in the region and globally, accelerating consumer demand by offering intuitive interfaces and outputs sometimes indistinguishable from human-generated content (OECD, 2023[1]).
These innovations present major opportunities for Asian capital markets. Investors and entrepreneurs are increasingly focused on identifying and investing in AI projects and start-ups with potential for significant market impacts, both in the financial sector and beyond. At the same time, managing the risks AI can pose is also crucial, to safeguard market stability and integrity, and to protect investors and financial consumers. This is especially important in a context where AI products are being deployed quickly, and where competitive dynamics could cause AI companies to prioritise speed over safety (OECD, 2024[2]). Policy tools such as innovation facilitators can play a crucial role in promoting responsible innovation to capture the benefits while managing risks.
AI-related merger and acquisition (M&A) activities, a proxy for the level of activity in this area, have remained elevated across Asia throughout 2024 and early 2025 (Figure 5.1, Panel A). Activity continued to be concentrated in China and Japan, together making up 65% of total M&A activity by value in 2024, followed by Korea, Hong Kong (China) and Singapore together making up 29%. M&A activities relating to the semiconductor industry – a critical component for AI model development – also remained high in 2024 and early 2025 (Panel B). Activity is heavily concentrated in China and Korea, which together made up 90% of total M&A activity by value in 2024, followed by Chinese Taipei and Hong Kong (China), which together made up 7%.
Elevated M&A deals illustrate the intensity of AI and semiconductor activity in the region
Note: In panel A, “Others” includes Indonesia, Malaysia, Chinese Taipei, Thailand and Viet Nam; and deal amount excludes semiconductors.
Source: LSEG; publicly available sources.
Investments in AI-related start-ups operating in the financial and insurance sectors have also remained elevated in recent years. The region saw a peak in the total value of venture capital (VC) deals in 2018, reaching USD 6.8 billion (129 deals), and a second peak of 205 VC deals in 2021, worth over USD 4.1 billion (Figure 5.2, Panel A). Broken down by economy, China dominates total VCs over the 2012-24 period, valued at over USD 14.2 billion (61.3% of total), followed by Indonesia at USD 4.4 billion (19.1% of total) and Singapore at USD 1.9 billion (8.4% of total) (Panel B).
China has led the region in VC investments into AI start-ups over the past 12 years
Note: In this figure, AI-related start-ups are defined as a private company that conducts research and delivers all or part of an AI system, or products and services that rely significantly on AI systems. The figure covers start-ups in the financial and insurance services sector. Panel A includes the annual total values for all economies shown in panel B, adjusted for inflation.
Source: OECD.AI visualisations powered by JSI using data from Preqin (2025[3]).
AI is boosting demand for data centre capacity across the region, building on the strong growth in supply since 2014
Note: This figure includes the following Southeast Asian jurisdictions: Indonesia, Malaysia, the Philippines, Singapore, Thailand and Viet Nam. The use of data centre capacity is not only limited to AI applications, and demand could be met by data centres both in the region and beyond.
Source: Rival et al. (2024[4]), Accelerating compute needs underpin Southeast Asia's rapid data centre growth.
Further, AI is driving a significant growth in demand for data centre capacity, especially across Asia (Figure 5.3). In Southeast Asia, total demand for data centre capacity is projected to roughly triple from 2023 to 2030, driven by a projected 10-fold increase in demand for AI compute. At the same time, live and pipeline supply1 across the Asia-Pacific region2 is expected to increase by 2.7 times between 2024 and 2028 (DC Byte, 2024[5]). Several hotspots for data centre investments have emerged, including across the Singapore – Johor (Malaysia) – Batam (Indonesia) corridor (Rival et al., 2024[4]). In early 2025, USD 2.7 billion worth of investments in data centres and cloud services were approved in Thailand (Reuters, 2025[6]). Other major cities in Asia that host large data centre clusters include Tokyo (2 561 megawatts), Mumbai (1 275 megawatts) and Seoul (1 254 megawatts) (DC Byte, 2024[5]).
The use of AI in finance can provide a range of benefits to markets and their participants, such as increased efficiency and accuracy of operational tasks, reduced costs, improved regulatory reporting and compliance, enhanced customer experiences, and more effective risk management (OECD, 2024[7]). AI innovation can also have benefits for emerging markets with large informal sectors, for example through strengthened fraud detection and applications in digital payment systems.
Innovation facilitators are an important tool for regulators and supervisors to proactively encourage AI and other types of innovation, while setting parameters to manage the potential risks of such innovation. This section maps the innovation facilitator arrangements deployed in Asian economies, and the extent to which these relate to AI innovations in finance as well as other sectors. Here innovation facilitators are defined as initiatives by or involving financial authorities that are designed to support innovation while limiting risks and safeguarding the financial systems (OECD, 2023[8]).
Regulators around the world are increasingly using novel or experimental approaches towards technological innovation (OECD, 2023[9]). While there is no globally agreed taxonomy of innovation facilitators, this section covers the most widely used types of innovation facilitators:
Sandboxes provide a temporary and limited regulatory waiver or flexibility to enable firms to test new business models with reduced requirements. Their main characteristics are that they are: (1) temporary; (2) use a trial-and-error approach; and (3) involve collaboration and iteration between stakeholders (OECD, 2023[9]).
Innovation hubs provide a dedicated point of contact for firms to interact with financial authorities on innovation-related matters, to navigate the regulatory/legal landscape and identify any specific issues or matters to consider. The interaction can include non-binding guidance on the conformity of innovative financial products, financial services or business models with licensing or registration requirements and regulatory and supervisory expectations (OECD, 2023[8]).
Regulatory accelerators enable partnership arrangements between innovators or FinTech firms and government authorities to ‘accelerate’ growth, innovate on shared technologies, and develop use cases that are particular to that authority. They are used to improve the familiarity of the regulator with FinTech products, concepts and firms (World Bank, 2020[10]).
Furthermore, there are a range of private sector led initiatives aiming to foster AI innovation. These include, for example, AI hubs, AI labs and AI accelerators. Services provided by these initiatives include dedicated workspaces for start-up companies, access to resources such as AI-compatible data centres, training courses and networking with established market participants. These initiatives are not covered in this chapter, which focuses on innovation facilitators established by financial authorities to complement market-based approaches to AI innovation.
Innovation facilitators are relatively recent arrangements, and first emerged in Western Europe and the United States, before being adopted more widely around the world. The first finance-related innovation hub was established in the Netherlands in 2007, and the first regulatory sandboxes were set up in the United States in 2012 and the United Kingdom in 2014 (IOSCO, 2022, p. 10[11]). Asian economies were also among the earliest adopters, with Japan introducing an innovation hub in 2015 and Hong Kong (China), Singapore and Malaysia all introducing sandboxes in 2016. The first innovation facilitator explicitly focused on AI was introduced by Singapore in 2023 and in 2024 Hong Kong (China) introduced the region’s first sandbox focused on AI in finance.
5.2.1. Implementation timeline across Asian economies
Innovation facilitators focused on AI or allowing for the testing of AI-related tools have existed in Asia for several years now. The examples provided in this Section showcase Asia’s continued leadership in developing innovation facilitators to promote AI-based innovation. Figure 5.4 provides a timeline of AI‑related innovation facilitators introduced across Asian jurisdictions, starting in 2016 with the Monetary Authority of Singapore’s (MAS) Regulatory Sandbox. MAS’s Sandbox aims to support innovation in a live but protected and time-restricted environment, including the possibility for MAS to temporarily relax certain legal and regulatory requirements (MAS, 2024[12]). While the MAS Sandbox is not focused on AI innovations, its scope enables AI use cases for the financial sector.
The Securities and Futures Commission (SFC) of Hong Kong also introduced in 2016 a “FinTech Contact Point”, allowing businesses intending to carry out regulated activities to seek information on regulatory requirements and potential licensing requirements (SFC, 2016[13]). This arrangement can be characterised as an innovation hub, as its objectives are to inform businesses seeking to innovate, and to ensure the regulator remains abreast of FinTech developments. The website of the FinTech Contact Point notes that the SFC welcomes discussions with firms that seek to introduce financial services based on emerging technologies, including AI.
In 2017, Korea’s AI Hub began to provide publicly available datasets to support AI model development while managing risks related to data quality and personal privacy. These include synthetic and real datasets in a range of formats and across a wide spectrum of fields, for example data to support facial recognition software tailored to the Korean market, and enhanced structured datasets drawn from Korean patent documents. Further, in 2024 the Korean AI Hub released a synthetic financial dataset assembled in partnership with several local data companies and Dong-eui University (AI Hub Korea, 2024[14]). The AI Hub also provides a range of other services, including access to high-performance computing resources and training programmes.
The first facilitator was introduced in 2016 and AI-focused facilitators have been established since 2023
Note: Only innovation facilitators with explicit AI aspects are included in this figure.
Source: Publicly available information (see Annex).
In 2022 Singapore introduced the Privacy Enhancing Technology (PET) Sandbox. Administered by the Infocomm Media Development Authority (IMDA), it aims to enable Singaporean companies to trial PET solutions and develop use cases by matching companies with approved PET providers, providing grant support to implement pilot projects, and providing regulatory support to ensure full compliance (IMDA Singapore, 2022[15]). The PET sandbox explicitly acknowledges the positive role that PETs can play in supporting cross-border data flows and responsible AI development. Accordingly, IMDA has encouraged businesses to participate in the sandbox to trial solutions relating to the use of PETs for training GenAI tools, or conversely the potential use of GenAI as a PET tool to identify and anonymise personal data (IMDA Singapore, 2024[16]).
In 2023, Singapore introduced a cross-sectoral GenAI Sandbox featuring major global technology firms such as Amazon Web Services, Google and Microsoft, consulting firms such as Deloitte and EY, AI companies such as Anthropic and Stability.AI, chipmaker NVIDIA and national companies such as Singtel and OCBC among others (IMDA Singapore, 2023[17]). The Sandbox’s objective is to support the development of trustworthy AI, by improving the evaluation of GenAI products and tools. It aims to map the evaluation benchmarks and methods for AI, to identify any gaps and to develop a common standardised framework.
This was followed in 2024 by the launch of the region’s first sandbox focused on AI in the financial sector, the Hong Kong Monetary Authority’s GenAI sandbox. Its objectives are similar to Singapore’s GenAI sandbox, namely to promote responsible GenAI innovation, with a focus on managing risk and engagement between companies and the regulator (HKMA, 2024[18]). It is however limited to the banking sector. The expected focus of eligible use cases include:
Risk management, such as optimising or automating credit assessments, generating risk assessment reports and strengthening risk warning systems.
Anti-fraud measures, such as the detection and prevention of AI-based scams, enhanced data analysis to identify fraudulent activities.
Improved customer experience, such as the use of advanced customer service chatbots providing customised responses.
A first cohort of 15 use cases was selected in December 2024 for technical trials (HKMA, 2024[19]). A second cohort was launched in April 2025 along with a new “GenAI Sandbox Collaboratory”, a series of practical workshops to facilitate engagement between banks and technology providers that may lead to future Sandbox trials (HKMA, 2025[20]).
In 2024, Malaysia’s Ministry of Science, Technology and Innovation (MOSTI) launched a dedicated AI Sandbox Pilot Programme (AI Programme), expanding Malaysia’s existing National Technology and Innovation Sandbox (NTIS), which covers innovations in several sectors and technologies. The AI Programme is administered by the Malaysian Research Accelerator for Technology and Innovation (MRANTI) in collaboration with NVIDIA Corporation, which will provide training, capacity building and technical support such as cloud credits for innovators and entrepreneurs. One of the AI Programme’s objectives is to facilitate the establishment of up to 900 domestic AI start-ups, as well as upskilling more than 13 000 Malaysians on AI technologies by 2026 (MRANTI Malaysia, 2024[21]). The AI Programme is open to participation by multiple sectors, including financial start-ups and innovators.
5.2.2. Key features and trends across the region
A detailed mapping of innovation facilitators across Asia is provided in the Annex. These include facilitators that explicitly aim to promote AI innovations (not only limited to the financial sector), that report past or current AI use cases in the financial sector, or that could in the future enable the promotion of AI innovations in the financial sector. Table 5.1 lists the 12 FinTech innovation facilitators with AI aspects identified in the region, including the HKMA’s GenAI Sandbox (the only facilitator focused on AI innovation in finance as mentioned above), general FinTech sandboxes, and innovation hubs that have reported AI use cases, one innovation hub featuring a specialised AI-based tool to combat illicit finance activities and two sandboxes where AI is explicitly referenced within the scope of eligible innovations.
Note: See also Table A.5 in the Annex for more information on these innovation facilitators.
Source: Publicly available information.
In addition to these government-led innovation facilitators, the Bank for International Settlements (BIS) operates a network of Innovation Hub Centres, with two Centres based in Hong Kong (China) and Singapore (BIS, 2021[22]). The Hong Kong (China) Centre focuses on Central Bank Digital Currency (CBDC) and green finance innovations, while the Singapore Centre focuses on suptech, regtech, CBDC and next-generation Financial Market Infrastructures. The BIS Innovation Hub has carried out a range of AI-related projects, with one project (“Project Symbiosis”) led by the Hong Kong (China) Centre. This project will apply AI and other emerging technologies to improve tracking and disclosure of the environmental impact of corporations and their suppliers. Project Symbiosis aims to encourage more sustainable practices (and funding for them) by providing better insights into the sustainability performance of a company’s supply chains (BIS, 2024[23]).
The mapping of innovation facilitators in Asia (Table 5.1) highlights several key features and trends. First, there is a limited number of AI-focused facilitators. As discussed, the only facilitator that explicitly focuses on AI innovation in the financial sector is the Hong Kong Monetary Authority’s (HKMA) GenAI sandbox for the banking sector. In addition, three facilitators in the financial sector include AI within the scope of technologies that are eligible, six facilitators have either reported at least one AI use case, one facilitator provides a dedicated AI solution to counter illicit finance, and one facilitator has commissioned research related to AI innovation. Beyond the financial sector, five innovation facilitators explicitly focus on AI innovation, and one facilitator includes AI among eligible innovations (see Annex for more details). Beyond these 18 facilitators with AI features, there is 20 general-purpose FinTech facilitators. A few innovation facilitators provide tailored services for AI development, including structured datasets for training models and access to AI-compatible computing power.
Second, most facilitators are designed to address specific sectors or products. Sandboxes and innovation hubs are generally administered by a single financial authority and apply only to innovations within a defined sector (for example banking) or product type (for example virtual assets). In some cases, general-purpose FinTech sandboxes operate on a cohort model, with the authority selecting different topics, products or sectors for each cohort. A few facilitators have been established at the sub-national level, for example the Seoul FinTech Lab (administered by the Metropolitan Government of Seoul), and the People’s Bank of China’s (PBOC) Fintech Innovation Regulatory Facility which has been deployed on a city-by-city basis. There are three exceptions to this approach. Cambodia’s FinTech Regulatory Sandbox is administered by three financial authorities under the purview of the Non-bank Financial Services Authority. The sandbox therefore covers insurance companies, trusts, and securities and exchange sectors. Second, India established in 2022 an Inter-operable Regulatory Sandbox, with the participation of five authorities covering the whole financial system. And the HKMA, PBOC and Monetary Authority of Macau have established a cross-border agreement to integrate their respective FinTech sandboxes.
Third, countries use different methods to communicate, ranging from formal public calls to informal engagement channels. Some provide regular public updates on the outcomes or decisions taken under their respective innovation facilitators, for example as case studies, lists of regulatory amendments or clarifications approved, or frequently asked questions (see section 5.3.1 for more details). The Securities and Exchange Commission of Pakistan (SECP) published a report in 2024 on experiences with its Regulatory Sandbox. On the other hand, some countries only announce the introduction of an innovation facilitator (or changes to its scope), without disclosing information on participation levels, outcomes or any decisions taken. While some jurisdictions publish detailed guidance on the objectives, scope, criteria and procedures of their facilitator, others only provide high-level information.
Facilitators also differ in the type and scope of support they offer - ranging from funding and tailored guidance to testing environments and international collaboration opportunities. For example, Singapore’s GenAI Sandbox for SMEs is developing a suite of GenAI solutions tailored to SMEs and selected SMEs will receive grant support to trial one solution of their choice for three months. Chinese Taipei’s FinTech Space offers counselling, mentoring and other services to support regulatory compliance, an on-site digital API environment linked to financial data for start-ups to trial solutions, and an international start-up exchange programme. Other facilitators provide more “traditional” services such as advice on regulatory requirements and testing of innovations within a sandbox environment.
Sandboxes are the most common type of innovation facilitators, accounting for 27 of the 38 analysed. There is some diversity in the scope and design of these sandboxes. Some allow for regulatory flexibility and potential relaxation of certain requirements, whereas others explicitly state that all requirements remain in force during the whole testing phase. Some are on a cohort basis, while others allow companies to apply at any time. There are different monitoring requirements, with some sandboxes requiring periodic self-reporting by participating entities, while others take a more intensive approach.
Reported experiences with innovation facilitators in Asia, some of which are provided below, showcase the benefits and challenges financial authorities face when using these facilitators to promote safe AI experimentation in the financial sector (Table 5.2 summarises some of the key benefits and challenges).
5.3.1. Reported experiences from innovation facilitators implemented in Asia
As previously discussed, countries have different practices for reporting on the implementation and use of innovation facilitators. Some countries publish the list of entities that have engaged with or used the facilitators. For example, Indonesia’s OJK reports the number of test cases in its Regulatory Sandbox, as well as the names of institutions that have successfully completed the sandbox. A 2024 report includes one company providing Machine Learning-based alternative credit scoring services, which received a “Recommendation” following participation in the sandbox (OJK, 2024[24]).
Other jurisdictions also report on the decisions and outcomes of each use case. For example, in Japan, the JFSA’s FinTech Proof-of-Concept (PoC) Hub reports specific compliance, supervisory or practical issues considered and decisions taken by the regulator, with regard to the interpretation of existing laws and regulations. In one PoC experiment, the JFSA confirmed that financial institutions may use AI tools to review records of interactions with consumers for the purpose of identifying potential breaches of requirements, so long as the reliability of the AI’s judgment criteria and trained models are verified in a reasonable manner and at reasonable intervals (Financial Services Agency, 2018[25]).
Another example is Korea’s FinTech Regulatory Sandbox website, which provides details on participating enterprises and outcomes of certain experiments (FinTech Center Korea, 2025[26]). These include a wide range of experiments involving GenAI and other AI-related applications. For example, one enterprise submitted in late-2024 multiple use cases involving both GenAI and other AI tools. One use case involved the use of internal AI tools to support information flows between employees, which triggered a review of regulatory requirements to separate the internal and external communication networks in financial companies. Several other use cases involved the use of a GenAI tool to provide personalised financial and trading analysis, coaching, consultation and related services to customers using both internal and external datasets. The website reports that “in case of granting an exception for network separation, restrictions are imposed on the range of permissible tasks and obligations to establish and implement security measures, in anticipation of potential information leakage and security breaches in the internal systems” (FinTech Center Korea, 2025[27]).
Some countries have also carried out self-evaluations on the experiences and effectiveness of their innovation facilitator arrangements. However, no country in the region has yet published a self-evaluation of an AI-focused facilitator. The SECP in Pakistan published a report in 2024 on its Regulatory Sandbox (SEC Pakistan, 2024[28]), covering the experiences and findings of the sandbox since its establishment in 2019. SECP reports that the sandbox has led to concrete regulatory changes, including the establishment of a framework to support peer-to-peer lending as well as micro and digital-only insurance companies. More specifically, the report notes that at least one AI-based algorithmic trading platform has been validated under the sandbox. In 2023, the sandbox switched from a cohort model to “always open” to allow more participation and a wider range of financial models, with a specific focus on conventional, Islamic and women-led entrepreneurs. Challenges were also reported, such as participants’ abilities to meet deadlines which were subsequently extended, and resourcing constraints for the SECP from a financial, staffing and expertise perspective.
Singapore’s IMDA has also published a series of case studies from use cases submitted to its PET Sandbox. One use case sets out the use of AI technologies to enhance customer engagement while preserving privacy. The use case, carried out by Ant International, involves a combination of Federated Learning, Multi-Party-Computing and Homomorphic Encryption to train a prediction model based on preference and behavioural data from their partners, without revealing sensitive information regarding their customers or business during the entire process (IMDA Singapore, n.d.[29]). The case study sets out the details of the use case, technical learnings and preliminary regulatory guidance. The conclusion of the case study states that Ant International intends to shift to live production and expects that similar tools are likely to be valuable in a range of business scenarios, such as fraud detection. The case study is published with the aim of providing demonstrations of the PET sandbox’ potential uses and the way it has already informed regulatory and business decisions to promote privacy-respecting innovations.
5.3.2. Benefits of innovation facilitators and their relevance for AI innovation
The fast-moving nature of AI means that regulatory frameworks need to be both adaptive and robust, encouraging innovation and testing while managing impacts and emerging risks such as potential market transformation. In an evolving and uncertain environment, regulators need to have channels to communicate with established and new market participants, and where relevant allow a certain level of experimentation to evaluate the impact of innovative technologies on business models, market dynamics and assess the effectiveness of existing policy frameworks. Such “experimentation” involves collecting evidence, identifying benefits and risks, and where necessary allowing new economic, institutional and technological approaches to be tested against the existing or new regulatory framework (OECD, 2024[30]). As discussed, innovation facilitators are one of the tools that can support these objectives, benefiting businesses, financial authorities and consumers.
Promoting responsible and safe innovation in financial markets is also crucial to support orderly capital market developments across Asia. Growth companies can benefit from tools aiming to encourage experimentation with the oversight and support of regulators. These companies, which are well-placed to introduce innovative technologies and business models based on AI and other technologies, can play an important role in stimulating Asian economies, but they also face challenges in accessing financing (OECD, 2024[31]).
Supporting businesses to deploy AI innovations in financial markets
By design, innovation facilitators seek to support testing and experimentation to encourage new and innovative products and services, within defined parameters to support other policy objectives, such as (depending on the case) consumer protection, personal data rights and anti-discrimination policies.
The first channel of support is by publicly signalling that authorities are prioritising and encouraging AI innovation (OECD, 2023[8]). Given the highly regulated nature of financial markets, such signals can support investor confidence in the viability of start-ups and other innovators seeking to introduce AI products, services and technologies. Signalling a pro-innovation approach can also contribute to market development by increasing entrepreneurs’ willingness to take risks, for example by establishing an AI-based start-up or investing an existing firm’s funds in an AI-related project. This signalling effect is however most effective in the context of an integrated innovation promotion strategy and potentially also a dedicated AI development strategy.
Second, innovation hubs or regulatory sandboxes provide an avenue to engage with the authorities, to seek clarification and to address potential shortcomings that would lead to regulatory incompatibility of an AI use case. These activities can contribute to legal certainty and a quicker path to market for companies (Attrey, Lesher and Lomax, 2020[32]).
Some jurisdictions also provide dedicated resources to support start-up businesses. For example, the HKMA GenAI Sandbox allows participating companies to access high-performance computers for AI development (HKMA, 2024[19]). Malaysia’s AI Sandbox Pilot Programme enables AI entrepreneurs and innovators to access specialised AI training, products and resources provided by the partnering entity NVIDIA (MRANTI Malaysia, 2024[21]).
While there is globally limited empirical information on the impacts of experimental regulatory approaches such as innovation facilitators, some evidence shows that they can help firms raise capital. A 2023 study of the United Kingdom Financial Conduct Authority’s regulatory sandbox found that participating firms saw a 15% increase in capital raised after joining the sandbox, while their probability of raising capital increased by 50% (Cornelli et al., 2020[33]). As detailed above, the signalling, regulatory and technical support channels can all contribute to increased attention by investors on specific AI use cases or innovative businesses that are being tested in such innovation facilitators.
Supporting competitive financial markets
Financial markets can benefit from efforts to support innovative firms seeking to apply AI and other technologies to provide new or improved products and services. Innovation facilitators can play an important role, especially those that provide dedicated resources or targeted support, such as access to high-performance computing power (for example the Hong Kong GenAI sandbox), AI-ready datasets to train models (for example the Korean Innovation Hub), financial grants, or access to expertise from established firms (for example the Malaysian AI Sandbox Programme in partnership with the private sector). This is particularly important for markets involving digital services, where incumbent BigTech companies or large financial institutions may have privileged access to customer data or other commercial advantages that may reduce start-ups’ ability to compete, depending on the market (OECD, 2021[34]). However, the design of innovation hubs could also pose challenges to the level playing field between companies, as discussed in 5.3.3 below.
The use of AI innovation facilitators can also foster the growth of capital markets by accelerating the adoption of AI innovation by market participants, benefiting from efficiencies and productivity gains associated with AI innovation and encouraging more dynamic and competitive capital markets. In addition, such initiatives can also promote financial inclusion, by allowing for new innovative products/services to be tested by smaller firms (e.g. FinTechs) and/or involving products and services designed for underbanked or unserved parts of the population.
Enhancing regulator’s awareness and understanding of AI innovations
In addition to improving market dynamics, innovation facilitators can also benefit participating authorities. By incentivising innovators to engage with the regulators, either through ad hoc queries or more intensive testing programmes, innovation hubs and sandboxes can improve the authorities’ awareness and understanding of market and technological developments. This can help to identify emerging risks to the financial system and develop appropriate policy responses (OECD, 2023[8]).
The amount and value of information collected can vary depending on the design of the facilitator and the amount of engagement from market participants. In any case, such information can provide an additional evidence base, for example by exposing specific barriers or gaps and demonstrating their tangible effects on innovation activities. The facilitator can also help to redress information asymmetries between the authorities and market participants (OECD, 2024[35]). This is particularly pertinent for advanced technologies such as AI, where collaborative approaches between regulators and market participants can help to better understand the impacts of new AI technologies, how they function, their intended objectives, and how best to measure and manage risks. These benefits can lead to improved regulatory design, more effective control and monitoring measures, and ultimately a virtuous cycle supporting further innovation.
Contributing to wider policy objectives
Innovation facilitators provide a safe space for assessing and testing AI innovations, which is crucial given the potentially significant risks AI can pose to financial markets. Risks include potentially biased and discriminatory outcomes and risks related to data quality and adequacy, the lack of explainability of models and challenges related to the robustness of the model outcomes, possible cyber-security and market manipulation risks, as well as risks related to the increased dependence on third-party service providers for AI-related services (OECD, 2024[30]; OECD, 2023[1]; OECD, 2021[36]). Testing can therefore support consumer protection objectives and improve consumer confidence by assessing the risks stemming from the use of such innovation.
For example, sandboxes provide a space to closely monitor a new AI product or service at smaller scale, allowing regulators to more quickly identify potential consumer protection concerns before widespread market deployment. They can provide a space for safe development or testing. For example, Korea’s AI Hub provides tailored synthetic (artificially generated) datasets for AI model testing, with the objective of reducing risks such as biases or misuse of personal data. Singapore’s PET Sandbox also provides spaces to test AI solutions that aim to improve personal data protection.
By promoting dialogue, innovation facilitators can also foster a culture of openness and trust between the financial authorities and market participants, with positive spillovers for all stakeholders (OECD, 2023[8]).
5.3.3. Challenges associated with innovation facilitators and their impacts for AI innovation
While innovation facilitators can be a powerful tool to promote innovation, potentially benefitting all stakeholders in the financial system, they also carry potential risks if not adequately designed and can lead to counterproductive outcomes. Furthermore, AI technologies have the potential for significant impacts across a wide range of stakeholders and touch on sensitive aspects such as privacy, access to employment and potential discrimination among others (OECD, 2024[37]). GenAI applications raise further potential risks, including more sophisticated fraud and cyber-attacks, reduced human oversight in decision-making and breaches of intellectual property rights (Lorenz, Perset and Berryhill, 2023[38]). This section focuses on the challenges that innovation facilitators face, with an emphasis on those related to AI in finance.
Potential ineffectiveness in promoting innovation
An innovation facilitator that is poorly designed, misaligned with market conditions, or not well-known and lacking engagement may fail to deliver meaningful benefits for AI innovation. A particular challenge for the promotion of AI innovation is the significant technical expertise regulators require to understand how these technologies function and the impacts they can have on financial markets. Such a lack of expertise can potentially prevent authorities from establishing innovation facilitators, or even lead to misleading conclusions if there is limited understanding of the impact of such innovation on financial activities. The lack of technical expertise could also lead to adverse outcomes, for example if a sandbox application is rejected because the regulator does not understand it, or if the sandbox experiment is rejected but might have succeeded if the company had gone directly to the market (OECD, 2023[9]). Furthermore, attracting technical experts to operate an innovation facilitator can be challenging, given the highly competitive labour market for AI-related engineering, science and technology experts.
Resourcing and opportunity costs
Innovation facilitators can require important levels of resourcing to operate successfully, with costs in the design, implementation and operational phases. These include direct costs such as data collection, analysis, stakeholder engagement, legal framework amendments (if required) and indirect costs such as opportunity costs (OECD, 2024[35]).
The amount of resourcing depends on the type of facilitator and its organisational arrangements. Sandboxes are generally considered to require greater financial and staffing investment to work effectively, while innovation hubs may be less investment intensive. At the same time, sandboxes may be more suitable for markets where there is already a developed and active innovation ecosystem with a variety of FinTech companies and existing products in development and/or deployed in the market (World Bank, 2020[10]). They are also considered to be more appropriate for companies with a high technological readiness level, with the ability to provide a robust, secure AI products and to comply with all safety and other regulatory requirements (OECD, 2023[8]). Inversely, innovation hubs may work best where a market is at an early stage of adopting AI innovations or for entities at an earlier stage of exploring potential AI use cases.
A 2018 study highlighted that many regulators had underestimated the resourcing required to establish and operate a sandbox (UNSGSA, 2018[39]). The level of resourcing required to operate an innovation facilitator can also determine its impact. It can affect the number of proposals or queries that can be processed, the time taken to provide responses or make decisions, the amount of communication with market participants, the quality and level of detail in the analysis of regulatory issues, and the capacity of companies to follow up. All these factors influence whether the innovation facilitator can effectively contribute to new products, services and technologies being introduced safely into financial markets. In other words, a poorly resourced facilitator may have sub-optimal or even counterproductive outcomes. For the same reasons, innovation facilitators may have a limited ability to scale up, which could limit the potential market-wide benefits (Attrey, Lesher and Lomax, 2020[32]). Innovation facilitators can also have indirect costs, most importantly opportunity costs.
Regulators will need to assess whether a facilitator is the optimal use of resources, by assessing the expected benefits of the arrangement (OECD, 2024[35]) and assessing the level of FinTech activity in the market to determine if other policy initiatives or reforms may be a better use of resources (World Bank, 2020[10]). Even where resourcing is available for a more intensive form of innovation facilitator such as a sandbox, regulators should still consider whether this approach is the most suitable based on the market conditions, challenges identified, alternative innovation promoting instruments available and opportunity cost (World Bank, 2020[10]).
As country experiences with innovation facilitators increase (both over time and with a greater number of facilitators introduced), the understanding of resourcing requirements and level of market interest will also improve. This highlights the importance of experience-sharing on common challenges (see Section 5.4)
Potential fragmentation of initiatives and outcomes
The establishment of innovation facilitators could accentuate risks of fragmentation or inconsistency depending on the design and co-operation arrangements between the different facilitators at the national, but also at the cross-border level. A significant degree of divergence can lead to regulatory arbitrage opportunities, where firms are treated differently according to the different innovation facilitators in place across countries or financial sub-sectors within the same country, which could lead to market distortions (OECD, 2023[9]). Such risks are greater where a facilitator considers relaxing certain regulatory requirements to promote innovation experimentation, particularly in the presence of competition to attract entrepreneurs and investments.
On the contrary, close co-operation of facilitators established by different authorities, or even the common establishment of jointly-operated facilitators at the national or regional level can promote greater consistency in the approaches and guidance offered to participating firms. The trilateral agreement between Hong Kong (China), China and Macau (China) for an integrated innovation facilitator framework is an example of authorities recognising the cross-border nature of FinTech innovations and the value of coordinated action (HKMA, 2024[40]).
The use of regulator-level innovation facilitators – while appropriate for innovations that are limited to a specific narrow sector under the competency of one authority – may be less effective for AI-related innovations with significant cross-industry applicability. The transversal nature of AI also means that multiple regulatory and legal regimes may apply to an innovative AI use case, for example on data protection, intellectual policy, consumer protection or other sector-specific laws (OECD, 2023[9]). Innovation facilitators with a narrow regulatory scope may be less effective in promoting AI innovation and may even increase complexity for participating firms. As an example, if regulatory guidance provided to a firm participating in a sector-specific innovation hub is not applicable to regulations in other sectors, or if the innovation hub staff do not have a broad understanding of all legal and regulatory aspects relevant to the proposed AI use case, the firm may not be able to scale its activities.
As mentioned above, some countries such as India have begun to address these issues by integrating or consolidating regulators’ sandboxes to allow a single point of entry for companies. In this case, each innovative use case submitted to the Inter-operable Regulatory Sandbox is assigned a ‘Principal Regulator’ responsible for assessing the application against its regulatory framework.3 The Principal Regulator can also invite other authorities to review and provide inputs, as ‘Associate Regulators’ (SEBI, 2022[41]). Five regulatory authorities are participating in the Inter-operable Regulatory Sandbox.
Impacts on the level playing-field between businesses
While innovation facilitators can contribute to innovations that spur competition, they can also create an unequal playing field between businesses. This risk is most acute for sandboxes, where a limited number of companies can apply to join a cohort of innovators. The participating businesses receive benefits from the regulators, such as dedicated guidance and advice, feedback on their products, and in some cases, special access to resources. This may lead to a competitive advantage compared to other firms that also seek to introduce innovative products and services (OECD, 2024[35]). Furthermore, businesses may use their participation in a sandbox or innovation hub as part of their branding, to imply an official endorsement or support from the regulators. This could lead consumers to favour such businesses compared to those excluded from the innovation facilitators.
These risks can be mitigated in several ways. Importantly, admission to the innovation facilitator should be based on objective and transparent criteria, with appropriate vetting processes (IOSCO, 2022[11]). Many sandbox operators also include disclaimers noting that participation in the sandbox does not constitute an endorsement of the business or its products.
Limited expertise among innovation facilitator staff
AI innovation involves sophisticated technologies that require expertise across a range of domains. These include statistics, computer programming, database management, data analysis and visualisation, cloud computing, specialised programming languages and machine learning (OECD, 2023[42]). Occupations related to AI also include mathematicians, actuaries and statisticians, software and application developers, ICT managers, database and network professionals, and electrotechnology engineers.
Innovative companies need to attract workers with these and other skillsets, including general cognitive and social skills considered both complementary and essential (OECD, 2023[42]), and there is a high level of competition in the labour market (Green and Lamby, 2023[43]). At the same time, authorities seeking to introduce innovation facilitators also need to attract workers with these skillsets, to ensure that they fully understand the AI use cases being tested. Without detailed technical knowledge, authorities’ understanding of the benefits, risks and implications of AI for finance may be limited. Furthermore, a lack of knowledge and understanding may also limit the effectiveness of innovation facilitators. For example, regulatory staff may not be able to provide adequate advice on whether a GenAI-based use case provides a sufficient level of explainability to meet regulatory requirements (OECD, 2023[1]). Singapore’s Privacy Enhancing Technology Sandbox provides an example of the importance of such technical expertise. The IMDA’s website includes case studies of specific AI tools (as discussed in section 5.3.1). These include detailed reporting on technical specifications and their implications for policy objectives such as protecting personal data. This example illustrates that facilitators may need to rely both on in-house AI expertise able to assess the proposed use cases and/or seek such expertise externally, with ensuing outsourcing implications (e.g. costs).4 Another example is Malaysia’s AI Sandbox Pilot Programme, which is a partnership between the Ministry of Science, Technology and Innovation and NVIDIA Corporation which provides AI expertise and training to innovative companies.
Potential risks to the reputation of financial authorities
There is a risk that regulatory authorities could suffer reputational damage if an innovation facilitator scheme is seen to be associated with business cases that ultimately cause investor, market or consumer harm (IOSCO, 2022[11]). This risk is especially high if the use of an innovation facilitator is seen as a label of quality, or a form of official endorsement by the regulator of the specific business case.
This section summarises the policy considerations for the design, operation and assessment of innovation facilitators to promote AI innovations in the financial sector across Asia. These policy considerations reflect the opportunities and challenges discussed in the previous section, as well as the current use of innovation facilitators in the region. They are also informed by relevant OECD standards, including the OECD Policy Framework for Effective and Efficient Financial Regulation (OECD, 2009[44]), the OECD AI Principles, (OECD, 2019[45]) and the OECD Recommendation for Agile Regulatory Governance to Harness Innovation (OECD, 2021[46]). Relevant provisions in these instruments are detailed in Annex.
Across Asian jurisdictions, there is a broad diversity in market sizes, levels of development, legal and regulatory frameworks, levels of international coordination, existing innovation ecosystems and approaches to financial regulation. However, authorities may consider the following issues, to ensure that responsible AI innovation can be effectively promoted.
Selecting the most appropriate innovation facilitator. Innovation facilitators can require a significant commitment of resources. Therefore, the choice should be aligned with a country’s market maturity, regulatory capacity and the technological readiness of participating firms. While sandboxes can offer in-depth testing environments, they typically require greater financial and staffing resources and are best suited to more developed innovation ecosystems. In contrast, innovation hubs may be more appropriate for markets at an earlier stage of AI adoption, providing flexible, lower-resource support for firms exploring emerging use cases.
Furthermore, different segments of the financial system within the same economy may have different levels of development in AI innovation. A sandbox may be effective for a market segment with a high level of innovation or expected future use cases, such as the financial sector, and may help to better understand the challenges and opportunities for that sector.
A related consideration is classifying which AI use cases are admissible to an innovation facilitator. A strictly sectoral approach could exclude innovations that have a tangle impact (providing either benefits or risks or both) on a specific sector or the financial system as a whole. It may be beneficial to include business cases where AI is used even if it is only indirectly related to finance. One of the guiding principles for policy makers could be whether the business case could have a material impact on the financial sector (OECD, 2023[8]).
Monitoring for continued relevance and effectiveness. Market conditions, technological innovations and the operation of innovation facilitators – including their outcomes – should be continuously monitored. This can involve collecting feedback from participating companies, documenting and if appropriate disclosing outcomes from the facilitators, assessing the level of awareness among market participants, and where necessary adjusting the facilitators to reflect evolving needs. At the same time, regulators should ensure that adjustments do not undermine objectives of regulatory certainty and fair treatment of participating entities. Changes can be communicated with market participants and industry bodies, with sufficient resourcing provided to respond to queries and provide clarification.
Ensuring collaboration and coordination between different innovation facilitators. The effectiveness of innovation facilitators may be enhanced by inter-agency co-operation frameworks and national-level strategies for AI development. Innovation facilitators at a sectoral level could be integrated, with a single point of contact for all market participants seeking regulatory clarification, wishing to test an AI use case, or seeking other types of assistance. Where this approach is not possible, regulators can enhance their information sharing and collaboration (OECD, 2023[8]), while considering communication strategies that help market participants navigate the different facilitators. India’s integrated sandbox framework is an example of such cross-sectoral coordination. Financial companies can apply via a single portal, with the potential use case being assessed and allocated to the most relevant participating authority (see section 5.3.3 for further detail).
Increasing international co-operation. The mapping of innovation facilitators across Asia has shown the wide variety of approaches taken in different jurisdictions. While sandboxes and innovation hubs are very widely used in the region, only a small number of these relate at least indirectly to AI in finance (see Section 5.2.2). Also, AI innovations are often transversal in nature, with an international supply chain in terms of software (e.g. cloud computing), hardware (data centres) and data (see Section 5.3.3). Furthermore, given the major social and economic impacts it can have, policy makers and regulators need a holistic approach to trustworthy and safe AI innovation (OECD, 2023[1]).
In this context, regulators would benefit from increased co-operation, to increase the potential for scalability of AI innovations, to enhance data collection and understanding of emerging AI risks, and to improve policy outcomes. The OECD AI Principles call for governments to actively co-operate to share knowledge on AI, build long-term expertise and develop global standards for interoperable AI (OECD, 2019[45]). Furthermore, the OECD Recommendation on Agile Regulatory Governance to Harness Innovation calls on adherents to strengthen co-operation between regulatory authorities at the sub-national, national and international levels to address the transboundary policy implications of innovation (OECD, 2021[46]). In this context, areas of co-operation for innovation facilitators could include:
Harmonisation of taxonomies: having an agreed set of definitions for different types of innovation facilitators could make it simpler for market participants to identify, navigate and use such tools. The mapping exercise in this chapter has highlighted a wide variety of names used for sandboxes and innovation hubs across jurisdictions (OECD, 2023[9]; IOSCO, 2022[11]).
Consistency across key criteria: increasing alignment in the criteria for different innovation facilitators – particularly on entry conditions and decision-making procedures – could contribute to a more level playing field between countries and between companies. In particular, it could help to avoid risks such as regulatory arbitrage or a race to the bottom where criteria are relaxed to attract innovators across borders (OECD, 2023[9]; OECD, 2024[35]). At the same time, the specificities and characteristic of the different countries in Asia should be taken into account for a more tailored approach where required.
Co-operation agreements between regulators: These can take the form of information sharing agreements, MoUs and reciprocal arrangements. Asian countries already have a network of such agreements, for example MAS’s FinTech Cooperation Agreements (MAS, 2023[47])5 and the agreement between Malaysia’s Securities Commission and Indonesia’s OJK. These could in the future have explicit references to, and terms for, innovation facilitator co-operation. These will contribute to improving knowledge and awareness of AI innovations, identifying potential risks, and improving some level of policy harmonisation to the extent possible.
Ensuring diverse AI expertise among facilitator staff. To ensure that innovation facilitators effectively contribute to promoting responsible AI innovation, regulators can consider building multidisciplinary teams with a mix of technical expertise and knowledge of emerging technologies. This would include retaining AI specialists who can advise on the technical aspects of a proposed product and service. To achieve this, regulators need to identify ways to retain experts, for example through opportunities to tackle more complex projects and contribute to knowledge creation (Guio, 2024[48]). In addition, the regulator can consider communication channels with industry experts or partner with AI champions. This is also in line with the OECD AI Principles which call for multi-stakeholder cooperation on AI (Principle 2.5) (OECD, 2019[45]).
References
[14] AI Hub Korea (2024), Financial Synthetic Data, https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=data&dataSetSn=71792 (accessed on 17 April 2025).
[32] Attrey, A., M. Lesher and C. Lomax (2020), “The role of sandboxes in promoting flexibility and innovation in the digital age”, OECD Going Digital Toolkit Notes, No. 2, OECD Publishing, Paris, https://doi.org/10.1787/cdf5ed45-en.
[23] BIS (2024), Project Symbiosis: AI and big data technologies for supply chain sustainability disclosure, https://www.bis.org/about/bisih/topics/cbdc/symbiosis.htm (accessed on 3 April 2025).
[22] BIS (2021), BIS Innovation Hub Centres, https://www.bis.org/about/bisih/locations.htm (accessed on 3 April 2025).
[33] Cornelli, G. et al. (2020), Regulatory sandboxes and fintech funding: evidence from the UK, Bank for International Settlements, http://www.bis.org (accessed on 12 May 2025).
[5] DC Byte (2024), Global Data Centre Index, https://www.dcbyte.com/2024-global-data-centre-index/ (accessed on 4 April 2025).
[25] Financial Services Agency (2018), Results of experiments for projects selected for support under the FinTech Demonstration Testing Hub, https://www.fsa.go.jp/news/30/20180801.html (accessed on 10 April 2025).
[26] FinTech Center Korea (2025), Business Partnership, https://sandbox.fintech.or.kr/business/enterprise_intro.do?lang=en&technologies=AI&pageIndex=3 (accessed on 10 April 2025).
[27] FinTech Center Korea (2025), Innovation financial service: KB Securities, https://sandbox.fintech.or.kr/business/enterprise.do?lang=en&id=523 (accessed on 6 May 2025).
[43] Green, A. and L. Lamby (2023), “The supply, demand and characteristics of the AI workforce across OECD countries”, OECD Social, Employment and Migration Working Papers, No. 287, OECD Publishing, Paris, https://doi.org/10.1787/bb17314a-en.
[48] Guio, A. (2024), “Regulatory sandboxes in developing economies: an innovative governance approach”, http://www.issuu.com/publicacionescepal/stacks (accessed on 28 March 2025).
[20] HKMA (2025), HKMA and Cyberport launch second cohort of GenA.I. Sandbox to accelerate A.I. innovation in financial sector, https://www.hkma.gov.hk/eng/news-and-media/press-releases/2025/04/20250428-5/ (accessed on 7 May 2025).
[40] HKMA (2024), “Expansion of Greater Bay Area Fintech Pilot Trial Facility”, https://cdn.amcm.gov.mo/uploads/attachment/2024-02/ch_av_614_mc006.pdf (accessed on 18 April 2025).
[18] HKMA (2024), HKMA and Cyberport Launch GenA.I. Sandbox to Bolster A.I. Adoption in Financial Sector, https://www.hkma.gov.hk/eng/news-and-media/press-releases/2024/08/20240813-6/ (accessed on 11 April 2025).
[19] HKMA (2024), HKMA announces inaugural cohort of GenA.I. Sandbox, https://www.hkma.gov.hk/eng/news-and-media/press-releases/2024/12/20241219-5/ (accessed on 11 April 2025).
[16] IMDA Singapore (2024), PETs Sandbox Expansion-Use of PETs for Generative AI Use cases, https://www.imda.gov.sg/-/media/imda/files/programme/pet-sandbox/pets-for-generative-ai.pdf (accessed on 18 April 2025).
[17] IMDA Singapore (2023), Generative AI Evaluation Sandbox, https://www.imda.gov.sg/resources/press-releases-factsheets-and-speeches/press-releases/2023/generative-ai-evaluation-sandbox (accessed on 11 April 2025).
[15] IMDA Singapore (2022), Privacy Enhancing Technology Sandboxes, https://www.imda.gov.sg/how-we-can-help/data-innovation/privacy-enhancing-technology-sandboxes (accessed on 18 April 2025).
[29] IMDA Singapore (n.d.), “Enhancing Customer Engagement with Privacy Preserving AI - ANT International Case Study”, https://www.imda.gov.sg/-/media/imda/files/programme/pet-sandbox/imda-pet-sandbox-case-study-ant-international.pdf (accessed on 18 April 2025).
[11] IOSCO (2022), The Use of Innovation Facilitators in Growth and Emerging Markets, http://www.iosco.org (accessed on 19 March 2025).
[38] Lorenz, P., K. Perset and J. Berryhill (2023), “Initial policy considerations for generative artificial intelligence”, OECD Artificial Intelligence Papers, No. 1, OECD Publishing, Paris, https://doi.org/10.1787/fae2d1e6-en.
[12] MAS (2024), Regulatory Sandbox, https://www.mas.gov.sg/development/fintech/regulatory-sandbox (accessed on 18 April 2025).
[47] MAS (2023), FinTech Cooperation Agreements, https://www.mas.gov.sg/development/fintech/fintech-cooperation-agreements (accessed on 28 April 2025).
[21] MRANTI Malaysia (2024), Mosti launches AI Sandbox Pilot Programme, https://mranti.my/happenings/news/mosti-launches-ai-sandbox-pilot-programme (accessed on 17 April 2025).
[3] OECD (2025), OECD.AI visualisations powered by JSI using data from Preqin, http://oecd.ai (accessed on 17 April 2025).
[37] OECD (2024), “AI, data governance and privacy: Synergies and areas of international co-operation”, OECD Artificial Intelligence Papers, No. 22, OECD Publishing, Paris, https://doi.org/10.1787/2476b1a4-en.
[2] OECD (2024), “Assessing potential future artificial intelligence risks, benefits and policy imperatives”, OECD Artificial Intelligence Papers, No. 27, OECD Publishing, Paris, https://doi.org/10.1787/3f4e3dfb-en.
[31] OECD (2024), Corporate Bond Markets in Asia: Challenges and Opportunities for Growth Companies, OECD Capital Market Series, OECD Publishing, Paris, https://doi.org/10.1787/96192f4a-en.
[7] OECD (2024), Mobilising ASEAN Capital Markets for Sustainable Growth, OECD Publishing, Paris, https://doi.org/10.1787/196b5bde-en.
[30] OECD (2024), “Regulatory approaches to Artificial Intelligence in finance”, OECD Artificial Intelligence Papers, No. 24, OECD Publishing, Paris, https://doi.org/10.1787/f1498c02-en.
[35] OECD (2024), “Regulatory experimentation: Moving ahead on the agile regulatory governance agenda”, OECD Public Governance Policy Papers, No. 47, OECD Publishing, Paris, https://doi.org/10.1787/f193910c-en.
[1] OECD (2023), “Generative artificial intelligence in finance”, OECD Artificial Intelligence Papers, No. 9, OECD Publishing, Paris, https://doi.org/10.1787/ac7149cc-en.
[42] OECD (2023), OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/08785bba-en.
[9] OECD (2023), “Regulatory sandboxes in artificial intelligence”, OECD Digital Economy Papers, No. 356, OECD Publishing, Paris, https://doi.org/10.1787/8f80a0e6-en.
[8] OECD (2023), Supporting FinTech Innovation in the Czech Republic: Regulatory Sandbox Design Considerations, OECD Publishing, Paris, https://doi.org/10.1787/081a005c-en.
[36] OECD (2021), Artificial Intelligence, Machine Learning and Big Data in Finance: Opportunities, Challenges and Implications for Policy Makers, OECD Publishing, Paris, https://doi.org/10.1787/98e761e7-en.
[34] OECD (2021), Digitalisation and Finance in Asia, OECD Publishing, Paris, https://doi.org/10.1787/dad32d34-en.
[46] OECD (2021), Recommendation of the Council for Agile Regulatory Governance to Harness Innovation, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0464.
[45] OECD (2019), AI Principles, https://www.oecd.org/en/topics/sub-issues/ai-principles.html (accessed on 29 April 2025).
[44] OECD (2009), Recommendation of the Council on a Policy Framework for Effective and Efficient Financial Regulation, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0377 (accessed on 21 March 2025).
[24] OJK (2024), Determination of Results of Regulatory Sandbox, https://ojk.go.id/en/fungsi-utama/itsk/regulatory-sandbox/default.aspx (accessed on 10 April 2025).
[6] Reuters (2025), Thailand approves $2.7 billion of investments in data centres, cloud services, https://www.reuters.com/technology/thailand-approves-27-billion-investments-data-centres-cloud-services-2025-03-17/ (accessed on 4 April 2025).
[4] Rival, O. et al. (2024), Accelerating compute needs underpin Southeast Asia’s rapid data center growth, https://web-assets.bcg.com/8b/c0/4ae607944bdd8d83068cfef87e31/accelerating-compute-needs-underpin-sea-dc-growth-vfinal.pdf.
[41] SEBI (2022), Standard Operating Procedure for Inter-operable Regulatory Sandbox, https://www.sebi.gov.in/media/press-releases/oct-2022/standard-operating-procedure-for-inter-operable-regulatory-sandbox-iors-_63948.html (accessed on 25 April 2025).
[28] SEC Pakistan (2024), Navigating Innovation: A Journey Through the Regulatory Sandbox, https://www.secp.gov.pk/document/rsb-report/?wpdmdl=54404&refresh=681a23197f7511746543385.
[13] SFC (2016), SFC establishes Fintech Contact Point, https://apps.sfc.hk/edistributionWeb/gateway/EN/news-and-announcements/news/doc?refNo=16PR19 (accessed on 23 April 2025).
[39] UNSGSA (2018), Early Lessons on Regulatory Innovations to Enable Inclusive FinTech: Innovation Offices, Regulatory Sandboxes, and RegTech.
[10] World Bank (2020), Global Experiences from Regulatory Sandboxes, https://documents1.worldbank.org/curated/en/912001605241080935/pdf/Global-Experiences-from-Regulatory-Sandboxes.pdf (accessed on 21 March 2025).
← 1. Defined as data centre capacity that is already operating, is committed, or is under construction.
← 2. Includes the following jurisdictions: Australia, Cambodia, Hong Kong (China), India, Indonesia, Japan, Korea, Malaysia, the Philippines, Singapore, Thailand, Chinese Taipei, Viet Nam and New Zealand.
← 3. The Principal Regulator is selected based on the ‘dominant feature’ of the use case, which falls under the remit of the given Regulator. The dominant feature is assessed based on the type of product involved, the level of impact on specific financial industries, and the number of regulatory relaxations sought by the applicant company.
← 4. InfoComm Media Development Authority of Singapore (IMDA) has not reported on staffing matters as part of the case studies published on the PET Sandbox website (IMDA Singapore, 2024[16]).
← 5. These include agreements with the National Bank of Cambodia, People's Bank of China, Hong Kong Monetary Authority (HKMA), Financial Services Authority of Indonesia (OJK), Japan Financial Services Agency (JFSA), Korean Financial Services Commission, Malaysia Securities Commission, Central Bank of the Philippines (BSP), Bank of Thailand (BOT) and State Bank of Vietnam.