全球智库信息集成服务系统

We identified and analysed a range of self-regulatory mechanisms — such as labelling initiatives, certification schemes, seals, trust/quality marks and codes of conduct — across diverse geographical contexts, sectors and AI applications.

The initiatives span different stages of development, from early stage (and still conceptual) proposed mechanisms to operational examples, but many have yet to gain widespread acceptance and use.

Many of the initiatives assess AI applications against ethical and legal criteria that emphasise safety, human rights and societal values, and are often based on principles that are informed by existing high-level ethical frameworks.

We found a series of opportunities and challenges associated with the design, development and implementation of these voluntary, self-regulatory tools for AI applications.

We outlined a set of key considerations that stakeholders can take forward to understand the potential implications for future action when designing, implementing and incentivising the take-up of voluntary, self-regulatory mechanisms, and to help contribute to the creation of a flexible and agile regulatory environment.

• Involving an independent and reputable organisation (for example, to carry out a third-party audit) could strengthen trust in an initiative, ensure effective oversight, and promote credibility and legitimacy.
• Actively engaging multiple interdisciplinary stakeholders to integrate a diversity of views and expertise in the design and development of AI self-regulatory tools could increase buy-in and adoption.
• The use of innovative approaches can help to address the perceived costs and burden associated with implementing self-regulatory mechanisms and also provide flexibility and adaptability in assessing AI systems.
• It is important to share learnings, communicate good practice, and for self-regulatory initiatives to be evaluated to track impacts and outcomes over time.
• There is a growing need for coordination and harmonisation of different initiatives to avoid the risk of a fragmented ecosystem and to promote clarity and understanding in the market.
• Rather than a one size fits all approach, it will be important to consider using a combination of different self-regulatory tools for diverse contexts and use cases to incentivise their voluntary adoption.

Artificial intelligence (AI) is recognised as a strategically important technology that can contribute to a wide array of societal and economic benefits. However, it is also a technology that may present serious challenges and have unintended consequences. Within this context, trust in AI is recognised as a key prerequisite for the broader uptake of this technology in society. It is therefore vital that AI products, services and systems are developed and implemented responsibly, safely and ethically.

Through a literature review, a crowdsourcing exercise and interviews with experts, we aimed to examine evidence on the use of labelling initiatives and schemes, codes of conduct and other voluntary, self-regulatory mechanisms for the ethical and safe development of AI applications. We draw out a set of common themes, highlight notable divergences between these mechanisms, and outline anticipated opportunities and challenges associated with developing and implementing them. We also offer a series of topics for further consideration to best balance these opportunities and challenges. These topics present a set of key learnings that stakeholders can take forward to understand the potential implications for future action when designing and implementing voluntary, self-regulatory mechanisms. The analysis is intended to stimulate further discussion and debate across stakeholders as applications of AI continue to multiply across the globe and particularly considering the European Commission's recently published draft proposal for AI regulation.

• Chapter One

  Introduction and overview

• Chapter Two

  The role of labelling initiatives, codes of conduct and other self-regulatory mechanisms in AI development and use

• Chapter Three

  Concluding remarks and reflections on the future

• Annex A

  Methodological approach

• Annex B

  Longlist of initiatives

• Annex C

  Detailed descriptions of some of the initiatives