In 'The Book of Why', Judea Pearl and Dana Mackenzie delve into the causal revolution, which has transformed the way we distinguish between correlation and causation. The book introduces causal diagrams, such as Directed Acyclic Graphs (DAGs), and explains how to predict the effects of interventions. It addresses fundamental questions about causality and its implications in fields like medicine, economics, and artificial intelligence. The authors also discuss the potential of causal inference in enabling computers to understand counterfactuals and engage in moral decision-making[2][4][5].
There are a few caveats to using generative AI tools, those caveats have led to a few tips that have quickly become second nature to those that use LLMs like ChatGPT. The main one being: have the domain knowledge to validate the output in order to avoid hallucinations. Hallucinations are one of the weak spots for LLMs due to the nature of the way they are built, as they are trained to correlate data in order to predict what might come next in an incomplete sequence. Does this mean that we’ll always have to be wary of the output of AI products, with the expectation that there is no intelligent decision-making going on under the hood? Far from it. Causal AI is bound by reason—rather than looking at correlation, these exciting systems are able to focus on the underlying causal mechanisms and relationships. As the AI field rapidly evolves, Causal AI is an area of research that is likely to have a huge impact on a huge number of industries and problems.
Paul Hünermund is an Assistant Professor of Strategy and Innovation at Copenhagen Business School. In his research, Dr. Hünermund studies how firms can leverage new technologies in the space of machine learning and artificial intelligence such as Causal AI for value creation and competitive advantage. His work explores the potential for biases in organizational decision-making and ways for managers to counter them. It thereby sheds light on the origins of effective business strategies in markets characterized by a high degree of technological competition and the resulting implications for economic growth and environmental sustainability.
His work has been published in The Journal of Management Studies, the Econometrics Journal, Research Policy, Journal of Product Innovation Management, International Journal of Industrial Organization, MIT Sloan Management Review, and Harvard Business Review, among others.
In the full episode, Richie and Paul explore Causal AI, its differences when compared to other forms of AI, use cases of Causal AI in fields like drug development, marketing, manufacturing, and defense. They also discuss how Causal AI contributes to better decision-making, the role of domain experts in getting accurate results, what happens in the early stages of Causal AI adoption, exciting new developments within the Causal AI space and much more.
Links Mentioned in the Show:
DataFramed 