世界经济论坛-人工智能治理之旅：发展与机遇（英）-2021.11-31页_3mb

AI Governance Journey Summary

Core Content

This report provides an overview of the evolution of AI governance, highlighting the journey from early skepticism to the current era of rapid development and increased regulatory focus. It outlines the key stages of AI governance, the challenges posed by AI's growth, and the importance of a multistakeholder and agile approach to ensure responsible AI use.

Main Points

1. Introduction

• The field of AI governance has grown significantly over the last five years.
• Responsible AI application and practices are critical to ensure that AI benefits and protects all of society.
• AI governance must balance innovation with oversight to prevent risks and maintain public trust.

2. Guidance on AI Governance

• AI governance involves more than just compliance and risk management; it includes strategy, market positioning, and competitiveness.
• Trust is essential in an AI-driven world and should be a key focus for companies.
• AI must be effective, safe, and deliver equitable results for all users.
• Cybersecurity is a critical concern as AI systems rely on large datasets.
• The stochastic nature of data-driven AI poses unique trust and risk challenges.
• AI has the potential to transform various sectors, including education, agriculture, and healthcare, but its governance remains a challenge due to the "pacing problem" – the mismatch between technological advancement and regulatory frameworks.

3. AI Governance Eras

• Pre-2010: AI Winter and Thaw
  Early AI development faced skepticism and limited progress, but there was always a mix of optimism and concern about its societal impact. Governance frameworks were not well established, and the focus was on balancing innovation and oversight.

• 2010-2016: AI Acceleration Presents Challenges
  AI began to accelerate due to increased computing power, cloud access, and data availability. This brought both opportunities and challenges, including ethical concerns and labor market disruptions. The "pacing problem" emerged as a key issue, highlighting the need for updated governance structures.

• 2016-2019: Principles and Guidelines
  Governments and industries started to develop AI ethics principles to guide development. These included privacy, fairness, transparency, and accountability. The Asilomar Conference in 2017 played a key role in raising awareness about AI ethics. Initiatives like the OpenAI Institute and the Partnership on AI were launched to promote responsible AI.

• 2019-present: Continued Acceleration and Governance Innovations
  The focus has shifted from principles to practice, with efforts to translate ethical guidelines into operational governance mechanisms. There is a growing recognition that existing regulations are insufficient, and new frameworks are being developed to address these gaps. The report highlights the importance of risk-based approaches, transparency, and public engagement in AI governance.

4. From Principles to Practice

• Over 100 ethical guidelines have been published by various stakeholders, showing progress in international cooperation.
• However, translating these into actionable governance mechanisms remains a challenge for many organizations.
• Companies are seeking internal governance structures and legal frameworks to manage AI risks without stifling innovation.

5. Multistakeholder Approaches

• Involving diverse stakeholders is essential for effective AI governance.
• Agile governance models are needed to adapt to the fast-paced nature of AI development.
• The Global AI Action Alliance is an initiative that promotes collaboration across sectors and borders.

6. The Road Ahead

• More tools and best practices are needed to support responsible AI governance.
• AI has the potential to impact jobs and inequality, requiring careful management.
• AI also contributes to carbon emissions, raising environmental concerns.
• Future governance gaps include the need for better understanding of AI's long-term implications and more robust frameworks to ensure ethical and equitable use.

7. Conclusion

• The Global AI Council plays a vital role in shaping international AI governance and cooperation.
• A multistakeholder and agile approach is necessary to address the unique challenges of AI.
• Continued innovation and collaboration are essential to ensure AI benefits all of society.

Key Information

• The Global AI Council is co-chaired by Brad Smith (Microsoft) and Kai Fu Lee (Sinovation Ventures), and includes leaders from government, business, academia, and civil society.
• The report highlights the "pacing problem", where laws and regulations struggle to keep up with AI's rapid development.
• Risk-based approaches are recommended to identify and prioritize governance gaps.
• Certification and labelling schemes are emerging as tools to promote ethical AI practices, with examples from Denmark, Malta, Singapore, and the IEEE.
• Public awareness and trust are crucial for the responsible adoption of AI, and initiatives like the Smart Toy Awards aim to celebrate ethical AI development.