Reproduce “Machine Culture” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Levin Brinkmann, Fabian Baumann, Jean-François Bonnefon, Maxime Derex, Thomas F. Müller, Anne-Marie Nussberger, Agnieszka Czaplicka, Alberto Acerbi and 7 more
“Machine Culture” frames artificial intelligence as an active participant in cultural evolution rather than solely a tool that affects human behavior. Using variation, transmission, and selection as organizing principles, the article connects generative models, reinforcement learning, language models, and recommender systems to a research agenda for measuring and governing human–machine cultural co-evolution.
Key insights
- AI can expand cultural variation through rapid search, nonhuman strategies, large-scale recombination, and semi-autonomous generation, while also introducing artifacts and ideas that may be difficult for humans to anticipate.
- Language models and other systems can preserve and transmit knowledge with high fidelity, but they may also reproduce biases, alter learning practices, and change which human skills remain valuable.
- Selection increasingly occurs through interacting human and machine processes: algorithms rank content and social connections, while human feedback, markets, and usage patterns shape AI behavior.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.