How Tyler Cowen and Alex Tabarrok Have Changed Their Views on AI and Where They Disagree

Tyler Cowen and Alex Tabarrok have both become more confident in the capabilities of artificial intelligence (AI) over time while maintaining an optimistic outlook on its economic potential. However, their views diverge on how these benefits will manifest, the risks associated with AI, and whether AI possesses consciousness.

Cowen emphasizes the role of institutions, bottlenecks, and geopolitical competition in the distribution of AI's benefits. He is wary of rapid technological advancement and its potential to exacerbate existing inequalities. Tabarrok, on the other hand, highlights inexpensive intelligence and scientific discovery as key drivers of progress. He also expresses growing concern about concrete safety failures, particularly in his more recent writings.

Their initial perspectives on the unequal distribution of gains from AI and the abundance it could create were shaped by Cowen's background in chess. He believed that strong computers required strong human partners, which would affect which skills commanded a premium. Tabarrok, in contrast, viewed AI as an increase in the world's supply of educated people, leading to more ideas and benefits that would be spread to consumers.

While neither Cowen nor Tabarrok predicted AI would completely replace human intelligence within a few decades, their views evolved significantly by 2023. By March of that year, Tabarrok explicitly endorsed Microsoft researchers' description of GPT-4 as exhibiting an incomplete form of general intelligence, marking a substantial shift from his previous stance.

Cowen reached a similar conclusion on April 16, 2025, acknowledging that AI could be AGI (Artificial General Intelligence), but emphasized the time needed to use it properly.

Despite their growing confidence in AI's capabilities, Cowen and Tabarrok remain cautious about the pace of adoption and the potential economic impact. In January 2025, Tabarrok predicted that the intelligence revolution would be bigger, more impactful, and more wrenching than the industrial revolution. However, Cowen was more tentative, suggesting that human bottlenecks would become more important as AI became more productive.

He estimated an additional half percentage point of annual growth, which, while significant when compounded, is far from an immediate economic explosion.

Both authors have also paid increasing attention to the potential risks and negative consequences of AI. Cowen has expressed concern about AI lowering net existential risks, while Tabarrok has focused on specific vulnerabilities, such as the combination of very powerful AI and its vulnerability to cybersecurity threats and manipulation. He emphasized that faster scientific advancements do not automatically translate to faster defense mechanisms, highlighting the importance of robust institutional safeguards.