Keynote by Mathias Cormann | OECD Secretary-General | India AI Impact Summit

Contents

Executive Summary

OECD Secretary-General Mathias Cormann outlined the transformative potential of AI while emphasizing that effective public policy is essential to enable responsible adoption and manage emerging risks. The keynote positioned the OECD as a central hub for evidence-based AI policymaking, presenting data on economic impacts, investment trends, incident tracking, and workforce transition frameworks that governments and businesses should consider when designing AI strategies.

Key Takeaways

AI productivity gains are significant but require active governance: The ~1 percentage point annual productivity boost depends on "strong adoption"—implying that policy choices fundamentally shape whether AI's economic benefits are realized.
U.S. capital dominance creates geopolitical asymmetries: The concentration of venture capital (75% of global AI deals) in the U.S. signals that non-U.S. countries need deliberate industrial and supply-chain strategies to avoid technological dependency.
The equity gap in AI readiness is already visible and widening: The 38-point gap in training participation between high and low literacy adults warns that without targeted intervention, AI adoption will exacerbate existing skills inequalities.
Incident tracking has moved from optional to essential: The tripling of reported AI incidents in three years demonstrates that structured, standardized incident reporting (via OECD frameworks) is now critical infrastructure for AI governance.
Multi-stakeholder coordination is necessary, not optional: The keynote repeatedly emphasizes that governments, industry, labor, and experts must work together—no single actor can manage AI's risks and opportunities alone.

Key Topics Covered

Economic Impact & Productivity: AI's potential to boost labor productivity across OECD and G20 economies
Global Investment Landscape: Venture capital concentration in AI startups and the geographic distribution of investment
Risk Tracking & Incident Reporting: Data on AI-related incidents and the need for standardized incident classification
Policy Frameworks & International Coordination: OECD tools for benchmarking AI policies and coordinating cross-national efforts
Workforce Transition & Skills: Job displacement risks and the equity gap in AI training access
Corporate Responsibility: Transparency frameworks and due diligence guidance for responsible AI innovation
Public-Private Collaboration: Need for integrated coordination between government, industry, labor, and experts

Key Points & Insights

Labor Productivity Gains: With strong adoption levels, AI could boost labor productivity by up to 1 percentage point annually across OECD and G20 countries over the next decade, translating to greater efficiency, lower costs, and higher living standards.
Massive Capital Investment: Nearly three-quarters of a trillion dollars in AI infrastructure investment is planned by major technology companies for a single year, underscoring the scale and pace of the technological shift.
Venture Capital Concentration: 61% of all global venture capital (259 billion USD) now flows to AI firms, up from 30% three years ago—with the U.S. capturing 75% of global AI venture capital deal value by a significant margin.
Exploding Incident Reports: AI-related incidents and hazards reported by media increased dramatically from 92 to 324 per month between 2022 and 2025, indicating growing safety and reliability concerns requiring systematic tracking.
Job Displacement Risk: Approximately 27% of employment is in occupations at the highest risk of automation, with critical equity gaps in skills training access—only 23% of adults with low literacy participate in AI training compared to 61% with higher literacy.
Agentic AI Adoption: Half of developers surveyed plan to integrate AI agents in their work, but security, privacy, and accuracy improvements are prerequisites for broader adoption.
Public Policy as Foundation: Historical analysis demonstrates that foundational AI technologies (internet, semiconductors) were shaped by public policy intervention, establishing the precedent for government's continued role in AI governance.
Policy Benchmarking Tools: The OECD released the AI Policy Index to help policymakers assess progress in implementing OECD AI recommendations, with an interactive toolkit featuring global best practices launching this year.
International Coordination Expansion: The Global Partnership on AI (GPI) has expanded to 46 countries across six continents, with Malta and Saudi Arabia joining, demonstrating broadening international consensus on AI governance.
Equitable Transition Framework: The Equitable AI Transitions Playbook (developed with the International Labour Organization) provides concrete policy examples for skills frameworks, upskilling, and reskilling to ensure inclusive workforce adaptation.

Notable Quotes or Statements

"With a strong level of adoption, AI could boost labor productivity by up to 1 percentage point every year across OECD and G20 countries over the next decade."

"The foundational technologies that made this technological revolution possible were very much shaped and supported by public policy—from internet connectivity to semiconductor supply chains."

"Between 2022 and 2025, in just three years, the number of AI incidents and hazards reported by the media increased dramatically from 92 to 324 per month on average."

"About 27% of employment is in occupations that are at the highest risk of automation. It will be particularly important to ensure access to training opportunities for those who need the most."

"To fully harness the enormous benefits and opportunities flowing from AI while mitigating and managing some of the associated risks and disruptions, we need to ensure governments, industry, labor, and experts work together to support responsible adoption."

Speakers & Organizations Mentioned

Entity	Role/Context
Mathias Cormann	OECD Secretary-General (primary speaker)
OECD	Primary institution providing analysis, frameworks, and policy guidance
India (host nation)	Leading the summit; demonstrates emerging-market engagement in AI governance
International Labour Organization (ILO)	Co-developer of the Equitable AI Transitions Playbook
Global Partnership on AI (GPI)	International coordination mechanism with 46 member countries; recently welcomed Malta and Saudi Arabia
Big Tech Companies (unnamed)	Collectively planning ~$750 billion in AI infrastructure investment
Hiroshima AI Process	Previous multilateral initiative; code of conduct framework launched at Paris AI Action Summit (2023)

Technical Concepts & Resources

Concept/Tool	Description
OECD AI Principles	Landmark foundational principles guiding responsible AI development and use
AI Policy Index	Evidence-based benchmark tool for assessing national AI policy implementation; released during this summit
Interactive AI Policy Toolkit	Launching this year; will feature repository of global AI policy best practices
OECD AI Incident Framework	Standardized reporting mechanism promoting global consistency in AI incident classification and reporting
Common Framework for AI Incident Reporting	System for tracking AI-related risks and hazards across jurisdictions
Agentic AI Landscape Report	Published the week before summit; identifies adoption trends and gaps in agent security, privacy, and accuracy
Hiroshima AI Process Code of Conduct (Updated)	Reporting framework for transparency and accountability; being revised to support SME adoption
OECD Due Diligence Guidance for Responsible AI	Published during summit; supports companies in navigating regulatory and voluntary frameworks
Equitable AI Transitions Playbook	Policy guide co-developed with ILO; provides concrete examples for skills frameworks, upskilling, and reskilling
Global AI Compute Capacity Tracking	OECD initiative tracking public AI compute distribution to inform national industrial strategy and supply-chain security
Global AI Investment Analysis	Ongoing OECD tracking showing venture capital allocation trends and geographic distribution

Note: The transcript contains notable repetition artifacts (transcription errors or speech stuttering), which have been filtered from this summary. The final segment transitions to a panel on "data sovereignty" with additional speakers announced but not yet speaking.