Broadening Economics in the Era of Artificial Intelligence and Experimental Evidence

The article addresses questions on how to form decisions, and how approaches founded on artificial intelligence can help us to improve them. It does so by discussing three exemplary case studies that are based on Niederreiter (Essays on contest experiments and supervised learning in the pharmaceutical industry, PhD thesis, IMT School for Advanced Studies Lucca, 2020) and complement this work. Each case study is a self-contained stream of work written such that different backgrounds, methodologies, and results are explained in sufficient depth to provide a base for future research. The first case study applies game theoretical learning models to laboratory data to understand how people learn in different competitive environments. The second case study uses a novel classification approach to identify latent behavioural types in such environments. The third case study employs a supervised learning method to obtain easily interpretable decision rules that aid at successfully classifying the outcome of clinical trials. Overall, the article advocates the importance of uniting approaches that originate outside mainstream economics but have the potential to broaden its portfolio and its appeal.