世界经济论坛-负责任地使用技术：IBM案例研究（英）-2021.9-28页_8mb

Responsible Use of Technology: The IBM Case Study Summary

Core Content

This document outlines IBM's approach to the responsible use of technology, focusing on the ethical development and deployment of artificial intelligence (AI) systems. It highlights IBM's commitment to trust, transparency, and accountability in the use of AI, emphasizing the need for integrating ethical considerations into the entire technology lifecycle. The document is part of a series published by the World Economic Forum to provide practical insights and tools for organizations aiming to adopt ethical AI practices.

Main Points

• Ethical Culture and Values: IBM has a long-standing commitment to ethics and privacy, which has been reinforced through a redefinition of company values in 2003. These values include dedication to client success, innovation that matters, and trust and personal responsibility.

• AI Ethics Board: Established in 2018, the AI Ethics Board plays a central role in guiding IBM's ethical AI initiatives. It is co-chaired by the Chief Privacy Officer and the Global Leader of AI Ethics, and includes representation from various business units and corporate functions.

• Local Focal Points and Advocacy Network: Each IBM business unit has a focal point to support the AI Ethics Board's work. The Advocacy Network of ethical technology champions promotes a culture of responsible AI within the company.

• Principles for Trust and Transparency: IBM has defined three key principles to guide its AI development:

  • The purpose of AI is to augment human intelligence
  • Data and insights belong to their creator
  • New technology, including AI systems, must be transparent and explainable

• Pillars of Trust: These are the foundational elements that support IBM's principles. The pillars include:

  • Explainability
  • Fairness
  • Robustness
  • Transparency
  • Privacy

• Toolkits and Research: IBM has developed several open-source and proprietary toolkits to support the implementation of these principles. These include:

  • AI Explainability 360: Provides tools for making AI models more transparent.
  • AI Fairness 360: Helps detect and mitigate bias in AI systems.
  • Adversarial Robustness 360: Offers defenses against adversarial attacks.
  • AI FactSheets 360: Enables transparency by providing detailed information about AI models.
  • Uncertainty Quantification 360: Helps assess the reliability of AI judgments.

• Proprietary Tools: IBM Cloud Pak for Data and Watson Studio are proprietary tools that support the ethical development and deployment of AI. These tools help in monitoring data quality, identifying bias, and managing AI models throughout their lifecycle.

Key Information

• IBM's approach to responsible AI includes both internal governance structures and external engagement with stakeholders.
• The AI Ethics Board was formed in response to growing concerns about AI ethics and has been instrumental in shaping IBM's ethical AI practices.
• The company has faced challenges, such as the 2018 Gender Shades study, which highlighted the need for ongoing ethical improvements and humility in AI development.
• IBM has developed a range of open-source and proprietary tools to support the implementation of its principles and pillars of trust.
• These initiatives are part of a broader effort to influence policy and integrate ethics into the AI development pipeline.

Structure and Implementation

• Governance Structure: The AI Ethics Board is supported by a Privacy Advisory Committee and a dedicated Project Office. It oversees the implementation of ethical AI practices across IBM.
• Education and Training: IBM has implemented a company-wide educational curriculum to promote responsible AI practices among employees.
• Collaboration and Engagement: IBM collaborates with a variety of stakeholders, including academia, government, and civil society, to advance ethical AI practices.
• Ethical Outcomes: IBM has developed capabilities such as "Trustworthy AI" and "Ethics by Design" to ensure ethical outcomes in AI development and use.

Conclusion

IBM's journey in the responsible use of technology reflects a commitment to ethical AI practices through a combination of internal governance, education, and collaboration. The company has made significant progress in operationalizing its ethical principles and has shared its insights to help other organizations follow a similar path. However, there is still work to be done in ensuring that ethical considerations are continuously integrated into the development and use of AI technologies.