AI Insights: The Tradeoff between Productivity and Cognition

An interesting dilemma is emerging in the AI literature. It’s clear that AI can deliver significant productivity gains. Studies show that knowledge workers experience research productivity gains of around 40%. An MIT study by Aidan Toner-Rodgers (2024) shows AI-assisted material scientists discovered 44% more materials, leading to a 39% increase in patent filings and a 17% rise in downstream product innovation. While Novy-Marx and Velikov (2025), showed how AI could be used to produce just under 200 finance papers using data from a large-scale event analysis study. The widely reported BCG/Harvard study showed that teams using GenAI for creative product innovation tasks outperformed their peer group by 40%.

The flip side is that this productivity revolution has complex cognitive implications. Lee et al. (2025) report the results of a survey of 319 knowledge workers exploring the impact of GenAI on critical thinking. A notable finding is the inverse correlation between AI trust and critical evaluation - as users develop higher confidence in AI systems, they demonstrate decreased critical thinking. The question this raises is whether skills will erode – if we deny people the opportunity to practice and hone their skills will they lose the ability (or never develop the skills they had). This is particularly important when you consider that the Toner-Rodgers study showed that more experienced people saw the biggest productivity gains when using AI, arguably because they were better able to decide which materials had potential and which did not.

It is clear the AI revolution presents significant productivity opportunities, but this is coupled with second order cognitive effects that need to be considered. Organisations need to think carefully about how to strategically deploy AI to achieve productivity gains, while maintaining cognitive sustainability. The fundamental question becomes not whether to adopt AI, but how to integrate these powerful tools while maintaining the essential human cognitive capabilities that drive sustainable innovation.

References:

Lee, H. et al (2025) The Impact of Generative AI on Critical Thinking, CHI, 2025.

Novy-Marx, R. and Velikov, M. (2025) AI-Powered (Finance) Scholarship,  January 8, available at: https://buff.ly/4hHGF2d.

Toner-Rodgers, A. (2024) Artificial Intelligence, Scientific Discovery, and Product Innovation, arXiv:2412.17866v1.

Postscript added 19/5/25 - MIT have announced that they no longer stand by the Toner-Rodgers study, so this should be taken into account when reading this blog - see https://lnkd.in/eKheRZCE