A new commentary published in the journal AI Perspectives argues that Alan Turing’s 1950 paper “Computing Machinery and Intelligence” anticipated modern machine learning through the idea that artificial intelligence would develop through a process resembling the education of a child, and that this vision raises an unsolved problem: how humans might teach or shape artificial intelligences to behave in ways that align with moral standards. The author, Stephen J. DeCanio, suggests part of such teaching would likely involve machines absorbing lessons from human writings, and demonstrates one small step using Latent Dirichlet Allocation (LDA), a “bag-of-words” statistical technique that analyzes texts by counting word frequencies while ignoring sentence syntax and can identify major themes without human intervention. LDA was applied to classical political theory texts downloaded in machine-readable form from Project Gutenberg. Results showed corpora made up of only a few authors and works could be successfully processed, and that natural language processing techniques are applicable to texts distant in time from the modern day. The author notes more computing power would enable results to encompass a wider range of texts and a greater number of topics in finer detail. All calculations were performed on a desktop PC running Mathematica 12. The author is emphatic that the findings do not, and cannot, demonstrate human-level understanding of texts by an unsupervised AI; the LDA “topics” are not summaries of the works’ contents.