Helen Toner (CSET): How to govern AI in the face of uncertainty?
Mar 13, 2024
auto_awesome
Helen Toner discusses China's AI policies, AI in warfare, challenges of AI regulation, and AI evaluations. They explore national security implications, integrating AI into daily life, and future AI technology like autonomous vehicles and energy advancements.
China actively regulates AI to balance security and innovation.
Smart regulation is crucial for AI industry growth and consumer trust.
Clear standards and governance structures are essential for AI evaluation and regulation.
Deep dives
Helen Toner's Journey Into AI Policy and National Security
Helen Toner, Director of Strategy at CSET and former Open Philanthropy senior research analyst, traces her interest in AI to the deep learning boom of 2011-2012. Encouraged by colleagues, she delved into the field, eventually focusing on AI policy at Open Philanthropy and China's AI regulations. Her research deciphers China's proactive stance on regulating AI, emphasizing security and innovation. Toner highlights the delicate balance China faces regarding social control and market growth, shedding light on AI's implications for both China and the US.
Regulating AI and National Security Implications in the US
Toner discusses the argument against AI regulation due to potential global AI race disadvantages, refuting it by showcasing China's active AI regulation. She emphasizes the need for smart regulation not hindering innovation, spotlighting the importance of consumer trust in driving industry growth. With a focus on national security, Toner emphasizes watching AI's impact on warfare and cybersecurity, stressing the need for careful deployment to avoid unintended consequences and maintain international dialogues.
Challenges in AI System Evaluation and Governance
The evaluation of AI systems poses significant challenges, with varying benchmarks and methodologies contributing to a currently chaotic evaluation landscape. Toner underscores the importance of clear standards for testing technologies, particularly for addressing issues like disparate impact. She also advocates for comprehensive governance structures to address uncertainties in regulating AI and proposes sector-specific regulators and new regulatory setups to deal with cross-cutting issues and emerging risks.
AI Agents and Future Concerns
Toner anticipates the rise of AI agents that carry out tasks autonomously, highlighting the potential for these agents to revolutionize daily life. She raises concerns about responsibility allocation and systemic risks as these agents operate in complex real-world environments. Toner also emphasizes the emerging vulnerabilities in AI systems interacting multifacetedly with the environment, underscoring the importance of addressing security and attack surface challenges to ensure the safe deployment of AI agents.
Exciting Future Innovations in AI Technology
Looking ahead, Toner expresses hope for transformative innovations in AI technology, particularly focusing on advancements in energy technology. From revolutionizing energy storage systems to optimizing energy allocation in grids, she envisions AI-driven breakthroughs in solar energy and nuclear fusion. Toner emphasizes the potential for AI to address key challenges across various sectors, signaling a future where efficient, AI-powered solutions drive positive change in energy and beyond.
This week Helen Toner, Director of Strategy and Foundational Research Grants at Georgetown’s Center for Security and Emerging Technology (CSET), joins In AI We Trust? to discuss decoding China’s AI policies, AI’s role in warfare, the potential impact of AI agents, challenges around regulating changing technology, and how to approach AI evaluations.