国际清算银行-人工智能和人力资本_央行面临的挑战（英）-2025.4_8页_501kb

BIS Bulletin Summary: Artificial Intelligence and Human Capital – Challenges for Central Banks

Core Content

This BIS Bulletin explores the impact of artificial intelligence (AI) on the human capital management of central banks, emphasizing the need for strategic adaptation in both recruitment and workforce development. It outlines two potential AI development scenarios and their implications for central bank operations and staffing.

Key Takeaways

• AI's Impact on Human Capital: AI is transforming how central banks utilize human capital, shifting job roles and increasing the demand for cross-disciplinary collaboration.
• Two Scenarios:
  • Scenario 1 ("Al Copilots"): AI tools based on large language models (LLMs) assist human experts, enhancing productivity without replacing human roles.
  • Scenario 2 ("Al Agents"): More autonomous AI agents could replace some human functions, requiring significant changes in workforce strategy and oversight.
• Workforce Adaptation: Central banks must focus on retraining and upskilling existing staff, attracting new talent, and fostering a culture of innovation to successfully integrate AI.

Introduction

AI is set to change the way central banks use human capital, affecting job roles and necessitating new approaches to talent management. The increasing reliance on AI for tasks such as real-time economic forecasting and transaction verification is leading to a more complex workforce planning landscape. Over 83% of central banks report increased complexity in workforce planning due to the integration of AI, and the limited supply of AI-savvy professionals exacerbates these challenges.

Scenario Descriptions

Scenario 1: LLM-based Copilot Systems

• AI tools assist human experts in daily tasks.
• Examples include internal chatbots, data analysis, transaction verification, and report generation.
• These tools are typically prompted with natural language and can produce text, code, images, and audio.
• AI enhances human capabilities, allowing staff to focus on more complex, high-level tasks.
• Staff need to be trained to use these tools effectively, with a focus on collaboration and interpreting AI insights.

Scenario 2: AI Agents

• Autonomous AI agents could replace some human roles with minimal oversight.
• These agents can perform tasks like real-time data gathering, forecasting, and transaction verification.
• They require more advanced technical skills, system management, and supervision.
• Governance frameworks must be embedded into the design of these agents to ensure ethical and legal compliance.
• New roles such as ML researchers, data engineers, and AI ethics officers may become more prevalent.

Workforce Implications

• Changing Job Profiles: New roles will be required to build and maintain AI applications, while existing roles will need additional skills.
• Work Practices: Continuous training and development will be essential for staff to adapt to evolving AI technologies and regulatory changes.
• Governance Frameworks: Central banks need clear ethical guidelines, data privacy standards, and accountability measures for AI deployment.
• Staffing Strategies: Central banks may need to rely on consultants, contractors, or outsourcing to address skill gaps, especially in scenario 2.
• Cultural Shifts: A culture of innovation and experimentation is necessary, requiring diverse teams and different working styles.

Human Capital Challenges

• Recruitment and Retention: Over 90% of central banks report increased difficulty in recruiting staff in the last five years.
• Skill Gaps: Central banks face challenges in hiring for roles such as cyber security, IT, fintech, data science, and AI/ML.
• Competitive Salaries: Public institutions may struggle to match private sector salaries for top AI talent.
• Career Progression: Perception of limited career opportunities may deter candidates, especially early-career professionals and those with technological expertise.
• Talent Management: Central banks must adopt multiple strategies to manage capability gaps and attract the right mix of skills.

Challenges in Recruitment and Retention

• Legal and Regulatory Restrictions: 58% of CBGN survey respondents face legal or regulatory constraints on recruitment.
• Sourcing and Citizenship Requirements: 47% cite sourcing requirements (e.g., civil service exams), and 43% mention citizenship requirements as barriers.
• Change Management: Central banks must implement effective change management strategies to address the transition to AI-intense workflows.
• Balanced Workforce Planning: Emphasizing development plans for both teams and individuals, with a balance between generalists and specialists, is crucial.
• Innovation Culture: Building diverse teams and fostering innovation is essential, though this may occur at a slower pace in central banks due to lower risk tolerance and the need for stakeholder engagement.

Conclusion

The integration of AI into central banking operations presents both opportunities and challenges for human capital management. Central banks must prepare for evolving job profiles, enhance staff skills, and develop robust governance frameworks to ensure responsible AI adoption. A balanced approach between permanent and agency staff is likely to be most effective in managing these transitions.