麦肯锡-生成式人工智能的经济潜力（英）-2023.6-68页_3mb

The report, "The Economic Potential of Generative AI: The Next Productivity Frontier," estimates that generative AI could add $2.6 trillion to $4.4 trillion annually to the global economy by 2030, potentially doubling the impact of traditional AI. This technology excels in augmenting knowledge work through use cases in customer operations, marketing, software engineering, and R&D, which collectively account for about 75% of the value. Industries like banking, high technology, and life sciences could see significant gains, such as $200 billion to $340 billion for banking. Key insights include accelerated automation of up to 50% of work activities by 2045, driven by improved natural language processing, which could boost labor productivity by 0.1% to 0.6% annually through 2040. However, challenges include managing risks like content reliability, intellectual property infringement, inadequate explainability, and workforce displacement. Businesses must address these risks through ethical deployment, retraining programs, and policy support to ensure sustainable growth and equitable outcomes.

Key Use Cases:

• Customer operations: Enhance service with AI chatbots and personalization.
• Marketing and sales: Automate content creation and customer engagement.
• Software engineering: Accelerate coding and testing.
• Product R&D: Enable faster drug discovery and design iterations.

Overall Economic and Workforce Impact:

• Automation could automate 50% to 60% of work activities by 2045, compared to 50% in previous estimates.
• Productivity growth is expected to increase due to higher output, with risks of widening inequality if workers are not supported.
• The technology's growth is spurred by rapid investment, with $12 billion invested in 2023 alone, primarily in North America.

Recommendations and Considerations:

• Companies should prioritize transparency, human oversight, and quality control in AI deployment.
• Policymakers and individuals need to adapt to workforce changes through retraining and education reforms.
• Risks like bias in generated content and security vulnerabilities require immediate attention to ensure responsible AI development.