Dan Williams on Why AI (Probably) Won’t Kill Us

In a recent conversation on The Good Fight Podcast, Yascha Mounk and Dan Williams, an Assistant Professor of Philosophy at the University of Sussex, explored the multifaceted implications of artificial intelligence (AI) on society. Williams emphasized the transformative potential of AI, comparing its impact to that of the Industrial Revolution, albeit occurring at a much faster pace. He argued that the ramifications of AI will be felt across various domains, including economics, culture, education, and the information environment. Williams suggested that addressing the complexities of AI requires an interdisciplinary approach, as the technology influences and intersects with numerous societal factors.

The discussion also touched on the political discourse surrounding AI, particularly the focus on issues like the construction of data centers. Mounk pointed out that many of the current political battles seem to miss the larger, more critical questions about AI’s societal implications, such as existential risks and the erosion of cognitive skills among younger generations. Williams concurred, suggesting that the political focus on data centers often reflects a NIMBY (Not In My Backyard) mentality rather than a genuine engagement with the significant changes AI is likely to bring.

The conversation highlighted the polarized views on AI—one camp perceiving it as merely another technology, while the other emphasizes the potential existential threats posed by superintelligent AI. Williams argued that both perspectives overlook the broader reality of AI’s transformative power. He acknowledged the importance of discussing existential risks but cautioned against viewing AI solely through that lens. He asserted that even aligned superintelligent AI could lead to a drastically different world, necessitating a more nuanced understanding of AI’s implications beyond binary categorizations of risk.

Williams and Mounk also referenced a widely discussed paper titled "AI as Normal Technology," which critiques the assumption that AI will seamlessly integrate into existing systems. The authors, Arvind Narayanan and Sayash Kapoor, argued that various real-world obstacles could hinder AI’s widespread adoption, despite its theoretical capabilities. This notion sparked a debate about whether AI can truly be considered a "normal" technology, with Williams suggesting that the existence of competing narratives about AI’s rollout further complicates the discussion.

Finally, Williams addressed concerns regarding the impact of AI on cognitive skills, particularly among younger generations. He expressed agreement with the worry that reliance on AI could undermine critical thinking and learning abilities. The conversation underscored the dual nature of AI as both a powerful educational tool and a potential barrier to genuine learning, emphasizing that the outcomes of AI integration into education will depend significantly on societal choices and institutional responses to the technology. This highlights the need for a thoughtful and proactive approach to harnessing AI’s potential while mitigating its risks.