Harnessing AI for Global Good by Gita Gopinath
Source details
- Canonical URL
- Harnessing AI for Global Good by Gita Gopinath
Other formats
Bibliographic details
- Authors: GITA GOPINATH
- Published: December 1, 2023
Key findings on productivity and labor markets
- Generative AI is spreading rapidly and could "unleash change on a similar scale" to past technological revolutions.
- Experiment with 453 college-educated professionals:
- Half given access to ChatGPT.
- Average time to complete writing assignments decreased by 40 percent.
- Quality of output rose by 18 percent.
- Firm-level studies indicate AI could raise annual labor productivity growth by:
- 2–3 percentage points on average.
- some studies show nearly 7 percentage points.
- Exposure to AI differs by country and occupation:
- Advanced economies (example: US and UK) have at least 60 percent of employment in high-exposure occupations.
- Emerging markets such as India: less than 30 percent of employment exposed to AI.
- Brazil and South Africa: closer to 40 percent exposure.
- Distributional effects:
- AI is expected to affect high-skill positions more than previous technologies that largely affected low-skill occupations.
- High-skill occupations may also gain more from complementary AI benefits, producing polarized outcomes within advanced economies.
- Long-term risks for developing economies:
- Labor-saving AI could re-shore tasks (example: call centers) and threaten traditional offshoring advantages, potentially complicating income convergence.
Ethical, societal, and financial system risks
- Generative AI’s mastery of language raises ethical concerns:
- Potential for "fake intimacy" to influence opinions and worldviews, with implications for social narratives and cultural norms.
- Prominent voices have warned of existential risks; more than 350 AI industry leaders signed a statement calling for priority on mitigating "extinction" risk from AI.
- Financial-sector risks:
- AI applications in risk assessments, credit underwriting, and investment recommendations could amplify herd behavior and stability risks if the financial system relies on only a few AI models.
- Lack of transparency in complex models will complicate post‑event analysis.
- Data privacy risks and embedded bias could exacerbate inequality (example: bots determining loan access).
- Research and development concentration:
- Until 2014, most machine learning models came from academia.
- In 2022, industry produced 32 significant machine learning models, compared with just three from academia.
- As state-of-the-art AI increasingly requires large data, computing power, and money, publicly funded AI research is advocated to highlight societal costs and maintain public interest research.
Public policy responses and global coordination
- Need for a coordinated global framework to maximize benefits and minimize harms; national approaches currently diverge:
- EU legislation classifies AI by risk levels.
- The EU’s approach differs from the US, which differs from the UK and China.
- Fragmented regulatory approaches risk slowing benefits and deepening geopolitical rivalries.
- Signs of progress toward coordination:
- G7’s Hiroshima AI process.
- U.S. executive order on AI.
- UK AI Safety Summit.
- UN call for a high-level advisory body on AI as part of its Global Digital Compact.
- Analogies and institutional models:
- The Paris Agreement offers a parallel for establishing shared frameworks.
- The Intergovernmental Panel on Climate Change is cited as a possible blueprint for a global expert group on AI.
- Policy priorities recommended:
- Develop global principles for responsible AI to harmonize local legislation.
- Design nimble social safety nets for workers displaced by AI.
- Reinvigorate labor market policies to keep workers attached to the labor market.
- Assess taxation policies to avoid favoring indiscriminate substitution of labor.
- Reform education systems to prepare new and current workers:
- Increase emphasis on STEM [science, technology, engineering, and math].
- Recognize potentially increased value of liberal arts education.
- Place education at the frontier of AI development and publicly fund AI research to balance industry concentration.
- Calibrate public instruments to support early-stage technologies without inducing fragmentation or protectionism.
- Support free flow of crucial resources—such as processors and data—and development of human and digital infrastructure.
Role of international financial institutions (IFIs)
- Three areas where IFIs can contribute:
- Knowledge gathering and analysis:
- IFIs can address broader economic and social effects of AI by gathering global-scale knowledge; IMF can use surveillance and internal expertise to explore AI challenges and opportunities.
- Convening and sharing best practices:
- IFIs can provide forums for sharing successful policy responses to build international consensus and harmonize regulations.
- Policy advice and supportive measures:
- IFIs can promote free flow of processors and data, and support the development of human and digital infrastructure.
- Policymakers should carefully calibrate public support to avoid fragmentation and protectionism; public investment in AI and related resources will remain necessary.
Conclusions and overarching message
- AI’s unique capacity to mimic human thinking requires a unique, globally coordinated set of rules and policies to ensure benefits for society.
- Multilateral cooperation is emphasized as more important than ever to avoid fragmented approaches and to pursue "responsible progress toward a better life for more people."
- The policy challenge is framed as potentially one of the most difficult public-policy tasks in recent history, requiring departure from echo chambers and consideration of humanity’s broad interest.
Source: "Harnessing AI for Global Good" by Gita Gopinath, F&D Magazine, December 2023.
Content in this bundle
- F&D: Harnessing AI for Global Good