Published Aug 17, 2024

Gary Marcus' keynote at AGI-24

Gary Marcus delivers a compelling keynote at AGI-24, arguing for stronger global and national regulations in AI to address ethical and economic challenges, critiques current scaling efforts, and shares insights on improving AI reliability and transparency.
Episode Highlights
Machine Learning Street Talk (MLST) logo

Popular Clips

Episode Highlights

  • Oversight

    underscores the necessity of layered oversight in AI development, drawing parallels to the aviation industry. He argues that independent auditors should have access to data to identify and address potential harms, emphasizing the importance of transparency and accountability 1. Marcus warns against regulatory capture, where big companies influence rules to their advantage, sidelining public interest 2.

    We need the government to make it possible for independent auditors to look and see what harms are actually taking place so that we can address them.

    ---

    He calls for the inclusion of independent scientists in regulatory discussions to prevent biased outcomes.

       

    AI Agencies

    Marcus proposes the creation of national and global AI agencies to oversee AI development and deployment. He criticizes the overhyped promises in the AI industry and highlights the gap between AI's potential and its current capabilities 3. He believes that such agencies could help manage AI's social impact and ensure ethical practices.

    Deep learning is a better ladder, but a better ladder doesn't necessarily get you to the moon.

    ---

    Marcus stresses the importance of these agencies in navigating the path to AGI and addressing the multifaceted risks associated with AI 4.

Related Episodes