Sandboxes for Artificial Intelligence

What are Regulatory Sandboxes (‘RS’)?

​​Regulatory sandboxes are a concept that emerged as a response to the challenges faced by governments relating to innovative technologies and businesses in regulated industries. Sandboxes are controlled environments where companies can test their innovative products, services, or business models under the supervision of regulatory authorities. Regulatory sandboxes aim to strike a balance between promoting innovation and ensuring consumer protection and regulatory compliance.

The UK Financial Conduct Authority (FCA) has defined it as “a ‘safe space’ in which businesses can test innovative products, services, business models and delivery mechanisms without immediately incurring all the normal regulatory consequences of engaging in the activity in question.”

In the context of Artificial Intelligence (AI), the proposed Artificial Intelligence Act by the European Parliament has defined sandboxes as “a tool allowing businesses to explore and experiment with new and innovative products, services or businesses under a regulator’s supervision.” Similarly, the Monetary Authority of Singapore defines it as a “framework [that] enables financial institutions and FinTech players to experiment with innovative financial products or services in a live environment but within a well-defined space and duration.”

What are the general benefits that RS can provide in practice?

  • Facilitating Innovation: businesses can test new technologies, business models, and products without being subject to the full weight of regulatory requirements initially. This can foster innovation by reducing barriers to entry and encouraging experimentation.
  • Collaboration and Learning: The exchange of information allows regulators to understand emerging technologies and business models better while companies gain valuable insights into regulatory expectations. Such collaboration can help regulators adapt their frameworks to accommodate technological advancements.
  • Regulatory Compliance: Despite providing a more flexible regulatory environment, regulatory sandboxes emphasize consumer protection and regulatory compliance. Companies must still adhere to certain safeguards and ensure customer data privacy, financial security, and other relevant regulations. The sandbox environment helps companies identify and address potential risks while regulators can assess the impact of innovative solutions on existing regulations.
  • Gradual Transition: Regulatory sandboxes typically involve a phased approach, allowing companies to gradually transition from the sandbox to full-scale operation.
  • Sector-Specific Focus:  By tailoring the sandbox to the unique challenges of each sector, regulators can address industry-specific concerns and create an environment conducive to innovation in those areas.

According to a World Bank study, more than 50 countries are currently experimenting with fintech sandboxes. Sandboxes can successfully add efficiency across different sectors. Singapore has been proactive in fostering innovation through its regulatory sandbox initiatives. The Monetary Authority of Singapore (MAS) introduced its fintech regulatory sandbox in 2016 and has expanded it to cover various sectors, including insurance, payments, and capital markets.

In the evolving AI industry, sandboxes can play a crucial role by offering a controlled environment or platform where developers and researchers can experiment, test, and refine their AI models, algorithms, or applications. 

Japan, Norway, and the United Kingdom have introduced sandboxes to provide a platform for experimenting with AI and other emerging technologies and understanding the associated challenges and opportunities.

China’s inaugural provincial-level AI legislation, known as the Shanghai Regulations on Promoting the Development of the AI Industry (the Shanghai AI Regulation), recommends the implementation of grading management and “sandbox” supervision, aiming to facilitate a favorable environment for companies to explore and test their technologies, thus contributing to the responsible and sustainable growth of the AI industry.

To summarize the benefits, regulatory sandboxes facilitate innovation by reducing barriers to entry, encouraging experimentation, and allowing businesses to test new technologies and models. They promote collaboration between regulators and companies by enabling a better understanding of emerging technologies and regulatory expectations. While providing a flexible environment, sandboxes prioritize consumer protection and regulatory compliance. 

It is important to also emphasise that Regulatory sandboxes can carry risks of being misused or abused, and many initiatives might find it challenging to strike the right balance between innovation and adequate safeguards for public and consumer protection. 

A report from the United Nations Secretary-General’s Special Advocate for Inclusive Finance for Development (UNSGSA) suggests that “around a quarter of regulators have launched sandbox initiatives without first evaluating feasibility, demand, potential outcomes, or collateral effects.” The report concludes that sandboxes are “not always the answer for regulating inclusive fintech.” 

It is crucial to prioritise the development of an adequate framework for analysis and preparation prior to launching a regulatory sandbox. Finally, by tailoring the regulations to specific sectors that address industry-specific challenges, regulatory sandboxes can support a gradual transition from the sandbox to full-scale operation.