麦肯锡：全球人工智能最新调研：AI在中国企业的落地进展如何？_英文版__21页_2mb

2022 AI State Summary: A Half-Decade in Review

Core Content

The 2022 McKinsey Global Survey on AI highlights the continued expansion of AI adoption across organizations, though growth has plateaued in recent years. Over the past five years, AI adoption has more than doubled, with 50% of organizations now using AI in at least one function, compared to 20% in 2017. The number of AI capabilities embedded in organizations has also doubled, from 1.9 in 2018 to 3.8 in 2022. Robotic process automation and computer vision remain the most commonly used AI capabilities, while natural-language text understanding has risen in prominence.

Main Points

AI Adoption and Use Cases

• Adoption Growth: AI adoption has more than doubled since 2017, with 50% of organizations now using AI.
• Stable Use Cases: Optimization of service operations has remained the top AI use case for the past four years.
• Functional Breakdown:
  • Marketing and sales: Top revenue contributors.
  • Product and service development: Also significant for revenue.
  • Supply chain management: Leading in cost benefits.
  • Risk and strategy: Increasingly important.
• Bottom-line Impact: About a quarter of respondents report that at least 5% of their EBIT was attributable to AI in 2021.

AI Investment and Spend

• Investment Increase: Over half of respondents now report spending more than 5% of their digital budgets on AI.
• Future Expectations: 63% of respondents expect AI investment to increase over the next three years.
• High performers lead in spend: AI high performers are nearly eight times more likely to spend at least 20% of their digital budgets on AI-related technologies.
• Enterprise Spend: High performers are over five times more likely to spend more than 20% of enterprise revenue on digital technologies.

AI and Sustainability

• AI in Sustainability: 43% of AI-adopting organizations use AI to support sustainability efforts.
• Environmental Impact: 40% of organizations are actively reducing AI-related emissions.
• Regional Differences: Greater China, Asia-Pacific, and developing markets are more likely to use AI for sustainability, while North America is least likely.
• Common Efforts: Energy efficiency, waste reduction, and regulatory compliance are the most common sustainability initiatives using AI.

AI Risk Mitigation

• No Substantial Progress: There has been no significant increase in the mitigation of AI-related risks since 2019.
• High performers lead in risk practices: AI high performers are more likely to engage in practices such as data governance, standardizing processes, and model testing.

Key Practices of AI High Performers

Strategic Alignment

• AI high performers are more likely to align their AI strategy with business outcomes.

Data and Model Management

• They often use modular data architectures and automate data processes.
• They are more likely to invest in AI-related technologies and use standardized tool sets.

Talent and Workforce Development

• High performers are more likely to engage nontechnical employees in AI development using low-code/no-code tools.
• They are more likely to hire roles like ML engineers and AI product managers.
• They also prioritize upskilling and reskilling, using experiential learning and certification programs.

Diversity and Inclusion

• Diversity Gaps: Women make up only 27% of AI team members on average, and racial/ethnic minorities make up 25%.
• Diversity Programs: 46% of organizations have active programs to increase gender diversity, while 33% focus on racial/ethnic diversity.
• Performance Correlation: Organizations with diverse AI teams are more likely to be high performers.

Challenges and Trends

• Talent Shortage: Hiring for AI roles remains challenging, with data scientists being particularly scarce.
• Reskilling: Reskilling and upskilling are common strategies to address talent gaps.
• Diversity Concerns: Despite efforts, diversity on AI teams remains a significant issue.
• AI Industrialization: Companies are moving towards industrializing AI, using practices like MLOps and standardized tool sets.

Conclusion

AI adoption has grown significantly over the past five years, with organizations increasingly embedding AI into their operations and processes. High performers are outpacing others in investment, talent acquisition, and best practices. However, challenges such as the talent shortage and diversity issues persist. The survey underscores the importance of strategic alignment, data governance, and workforce development in achieving AI success. While AI continues to deliver bottom-line value, the lack of progress in risk mitigation and diversity remains a concern. Organizations outside the high performer group can benefit from adopting the best practices observed in leaders to improve their AI capabilities.