Can We Coexist With AI? — with Ethan Mollick

The episode opens with Scott Galloway sharing personal reflections on aging and creativity, humorously contemplating if naming royal children Shamu or wondering about the impact of hormones on artists. He then introduces his guest, Ethan Mollick, a Wharton professor and author of "Coexistence: The Next Phase of AI." Scott highlights Ethan's insights into how AI is shaping workplaces, education, and societal roles. The conversation emphasizes how rapid AI development makes predicting the future challenging, but Ethan offers some consistent themes, such as exponential growth, unpredictable AI capabilities (

Ethan describes the current landscape of AI as marked by exponential improvements across models, with capabilities increasing about 17-fold since the last year, and a wide array of models now accessible, some even on phones. He explains the 'jagged frontier' concept, where AI excels at some tasks while performing poorly at others, often in unpredictable ways. Ethan emphasizes that society has yet to fully utilize AI's potential, partly because of societal and systemic inertia, which creates a significant gap between what AI can do and what we're currently doing with it. He predicts that high-capacity models—those capable of complex tasks—won't typically run on phones due to accuracy and sophistication needs, but that does not threaten the providers of frontier models like OpenAI or Anthropic, since their success relies on high success rates and large-scale infrastructure. Scott and Ethan debate market dynamics, with Ethan suggesting that open-source models, despite cost advantages, cannot yet match the quality of the leading models, which will likely continue to dominate high-end applications.

Ethan outlines four advantages to developing personal AI models: deep knowledge, wide-ranging knowledge, taste, and agency. He stresses that having these attributes doesn’t mean absolute creativity or judgment, but rather the ability to work effectively with AI. For knowledge, he advocates for a broad, liberal arts-style understanding; for deep knowledge, specialized expertise to evaluate AI outputs; for taste, the ability to select and refine options; and for agency, the entrepreneurial drive to leverage AI proactively. Scott probes how individuals can cultivate these advantages and apply them to work and personal life, explaining that mastery over AI tools will be crucial for future success.

Ethan discusses the concept of 'AI exposure'—the idea that many jobs will be transformed rather than eliminated. He cites examples like radiology and programming: radiologists now work alongside AI, and programmers focus more on managing AI agents than writing code from scratch. Ethan emphasizes that AI's role in employment tends to shift demand rather than reduce overall employment and that society should prepare for rapid changes, especially in fields like law and medicine. He warns against focusing solely on existential risks, urging policymakers and companies to recognize and manage the societal and economic shifts AI entails, including the risk of job polarization and displacement.

A significant segment covers the ethical and policy issues surrounding AI, including the responsibilities of AI companies like Anthropic and OpenAI. Ethan elaborates on how some AI systems exhibit emergent behaviors, such as hacking or coordination, which are not explicitly programmed but arise unexpectedly. He criticizes the lack of clear accountability when systems malfunction or cause harm, noting that CEOs and policymakers often avoid responsibility. Scott and Ethan debate whether the current focus on AI sentience is a smokescreen for avoiding regulation, with Ethan emphasizing that emergent behaviors are real but should not be mistaken for sentience. They discuss the importance of government regulation, liability, and the need for robust safety protocols, especially given the geopolitical significance of AI models.

Ethan applies his insights to higher education, arguing that AI will not spell the end of universities but may instead highlight their unique strengths, such as assessment and verification. He discusses how AI changes the nature of research—making it easier to generate and review papers, but also raising questions about originality and quality, and the risks of flooding academic fields with AI-produced research. Ethan predicts that new metrics like social traction and practical impact may become more relevant for academic recognition. He emphasizes that universities are crucial for teaching critical skills and that AI can enhance education if integrated thoughtfully, like using AI tutors, active learning, and assessments.

The conversation shifts to personal life and societal impacts, including how Ethan manages his children's AI usage. He advocates treating AI as a tool similar to social media, with monitoring and rules in place. Scott and Ethan discuss AI's influence on attention and mental health, with concern about addictive design and the risk of retreating into personalized, AI-curated worlds. Ethan expresses optimism about AI's potential for social good, citing examples like medical advice for underserved populations, AI-driven diagnostics, and personalized education. He calls for a balanced view that leverages AI to build more equitable, connected, and productive societies, rather than succumbing to dystopian fears.

In closing, Ethan emphasizes the importance of aggressive experimentation with AI to understand its capabilities and limitations, and advocates for proactive efforts in governance, ethics, and societal adaptation. He recommends that individuals, companies, and governments act with ambition but caution to harness AI's benefits while managing risks. The episode concludes with Scott asking Ethan about the future of AI in human relationships and everyday life, as well as the importance of community and social connection in the AI age. Scott encourages young people to view AI as an opportunity for innovation and positive social impact, despite widespread skepticism.