Cultural evolution allows ideas and technology to build over generations, a process reaching its most complex and open-ended form in humans. While social learning enables the transmission of such innovations, the cognitive processes that generate innovations remain unclear. We propose that semantic knowledge-the associations linking concepts to their properties and functions-guides human innovation and drives cumulative culture. To test this, we combined an agent-based model, which examines how semantic knowledge shapes cultural evolutionary dynamics, with a large-scale behavioural experiment (N = 1,243) testing its role in human innovation. Semantic knowledge directed exploration toward meaningful solutions and interacted synergistically with social learning to amplify innovation and cultural evolution. Participants lacking access to semantic knowledge performed no better than chance, even when social information was available, and relied on shallow exploration strategies for innovation. Together, these findings indicate that semantic knowledge is a key cognitive process enabling human cumulative culture.

The convergence of humans and artificial intelligence systems introduces new dynamics into the cultural and intellectual landscape. Complementing emerging cultural evolution concepts such as machine culture, AI agents represent a significant techno-sociological development, particularly within the anthropological study of Web3 as a community focused on decentralization through blockchain. Despite their growing presence, the cultural significance of AI agents remains largely unexplored in academic literature. Toward this end, we conceived hybrid netnography, a novel interdisciplinary approach that examines the cultural and intellectual dynamics within digital ecosystems by analyzing the interactions and contributions of both human and AI agents as co-participants in shaping narratives, ideas, and cultural artifacts. We argue that, within the Web3 community on the social media platform X, these agents challenge traditional notions of participation and influence in public discourse, creating a hybrid marketplace of ideas, a conceptual space where human and AI generated ideas coexist and compete for attention. We examine the current state of AI agents in idea generation, propagation, and engagement, positioning their role as cultural agents through the lens of memetics and encouraging further inquiry into their cultural and societal impact. Additionally, we address the implications of this paradigm for privacy, intellectual property, and governance, highlighting the societal and legal challenges of integrating AI agents into the hybrid marketplace of ideas.

Large language models (LLMs) provide a compelling foundation for building generally-capable AI agents. These agents may soon be deployed at scale in the real world, representing the interests of individual humans (e.g., AI assistants) or groups of humans (e.g., AI-accelerated corporations). At present, relatively little is known about the dynamics of multiple LLM agents interacting over many generations of iterative deployment. In this paper, we examine whether a "society" of LLM agents can learn mutually beneficial social norms in the face of incentives to defect, a distinctive feature of human sociality that is arguably crucial to the success of civilization. In particular, we study the evolution of indirect reciprocity across generations of LLM agents playing a classic iterated Donor Game in which agents can observe the recent behavior of their peers. We find that the evolution of cooperation differs markedly across base models, with societies of Claude 3.5 Sonnet agents achieving significantly higher average scores than Gemini 1.5 Flash, which, in turn, outperforms GPT-4o. Further, Claude 3.5 Sonnet can make use of an additional mechanism for costly punishment to achieve yet higher scores, while Gemini 1.5 Flash and GPT-4o fail to do so. For each model class, we also observe variation in emergent behavior across random seeds, suggesting an understudied sensitive dependence on initial conditions. We suggest that our evaluation regime could inspire an inexpensive and informative new class of LLM benchmarks, focussed on the implications of LLM agent deployment for the cooperative infrastructure of society.

Human culture relies on collective innovation: our ability to continuously explore how existing elements in our environment can be combined to create new ones. Language is hypothesized to play a key role in human culture, driving individual cognitive capacities and shaping communication. Yet the majority of models of collective innovation assign no cognitive capacities or language abilities to agents. Here, we contribute a computational study of collective innovation where agents are Large Language Models (LLMs) that play Little Alchemy 2, a creative video game originally developed for humans that, as we argue, captures useful aspects of innovation landscapes not present in previous test-beds. We, first, study an LLM in isolation and discover that it exhibits both useful skills and crucial limitations. We, then, study groups of LLMs that share information related to their behaviour and focus on the effect of social connectivity on collective performance. In agreement with previous human and computational studies, we observe that groups with dynamic connectivity out-compete fully-connected groups. Our work reveals opportunities and challenges for future studies of collective innovation that are becoming increasingly relevant as Generative Artificial Intelligence algorithms and humans innovate alongside each other.

As large language models (LLMs) start interacting with each other and generating an increasing amount of text online, it becomes crucial to better understand how information is transformed as it passes from one LLM to the next. While significant research has examined individual LLM behaviors, existing studies have largely overlooked the collective behaviors and information distortions arising from iterated LLM interactions. Small biases, negligible at the single output level, risk being amplified in iterated interactions, potentially leading the content to evolve towards attractor states. In a series of telephone game experiments, we apply a transmission chain design borrowed from the human cultural evolution literature: LLM agents iteratively receive, produce, and transmit texts from the previous to the next agent in the chain. By tracking the evolution of text toxicity, positivity, difficulty, and length across transmission chains, we uncover the existence of biases and attractors, and study their dependence on the initial text, the instructions, language model, and model size. For instance, we find that more open-ended instructions lead to stronger attraction effects compared to more constrained tasks. We also find that different text properties display different sensitivity to attraction effects, with toxicity leading to stronger attractors than length. These findings highlight the importance of accounting for multi-step transmission dynamics and represent a first step towards a more comprehensive understanding of LLM cultural dynamics.

This paper is an invited layperson summary for The Academic of the paper referenced on the last page. We summarize how the formal framework of autocatalytic networks offers a means of modeling the origins of self-organizing, self-sustaining structures that are sufficiently complex to reproduce and evolve, be they organisms undergoing biological evolution, novelty-generating minds driving cultural evolution, or artificial intelligence networks such as large language models. The approach can be used to analyze and detect phase transitions in vastly complex networks that have proven intractable with other approaches, and suggests a promising avenue to building an autonomous, agentic AI self. It seems reasonable to expect that such an autocatalytic AI would possess creative agency akin to that of humans, and undergo psychologically healing -- i.e., therapeutic -- internal transformation through engagement in creative tasks. Moreover, creative tasks would be expected to help such an AI solidify its self-identity.

Research in cultural evolution aims at providing causal explanations for the change of culture over time. Over the past decades, this field has generated an important body of knowledge, using experimental, historical, and computational methods. While computational models have been very successful at generating testable hypotheses about the effects of several factors, such as population structure or transmission biases, some phenomena have so far been more complex to capture using agent-based and formal models. This is in particular the case for the effect of the transformations of social information induced by evolved cognitive mechanisms. We here propose that leveraging the capacity of Large Language Models (LLMs) to mimic human behavior may be fruitful to address this gap. On top of being an useful approximation of human cultural dynamics, multi-agents models featuring generative agents are also important to study for their own sake. Indeed, as artificial agents are bound to participate more and more to the evolution of culture, it is crucial to better understand the dynamics of machine-generated cultural evolution. We here present a framework for simulating cultural evolution in populations of LLMs, allowing the manipulation of variables known to be important in cultural evolution, such as network structure, personality, and the way social information is aggregated and transformed. The software we developed for conducting these simulations is open-source and features an intuitive user-interface, which we hope will help to build bridges between the fields of cultural evolution and generative artificial intelligence.

The goal of Artificial Life research, as articulated by Chris Langton, is "to contribute to theoretical biology by locating life-as-we-know-it within the larger picture of life-as-it-could-be" (1989, p.1). The study and pursuit of open-ended evolution in artificial evolutionary systems exemplifies this goal. However, open-ended evolution research is hampered by two fundamental issues; the struggle to replicate open-endedness in an artificial evolutionary system, and the fact that we only have one system (genetic evolution) from which to draw inspiration. Here we argue that cultural evolution should be seen not only as another real-world example of an open-ended evolutionary system, but that the unique qualities seen in cultural evolution provide us with a new perspective from which we can assess the fundamental properties of, and ask new questions about, open-ended evolutionary systems, especially in regard to evolved open-endedness and transitions from bounded to unbounded evolution. Here we provide an overview of culture as an evolutionary system, highlight the interesting case of human cultural evolution as an open-ended evolutionary system, and contextualise cultural evolution under the framework of (evolved) open-ended evolution. We go on to provide a set of new questions that can be asked once we consider cultural evolution within the framework of open-ended evolution, and introduce new insights that we may be able to gain about evolved open-endedness as a result of asking these questions.

Knowledge built culturally across generations allows humans to learn far more than an individual could glean from their own experience in a lifetime. Cultural knowledge in turn rests on language: language is the richest record of what previous generations believed, valued, and practiced. The power and mechanisms of language as a means of cultural learning, however, are not well understood. We take a first step towards reverse-engineering cultural learning through language. We developed a suite of complex high-stakes tasks in the form of minimalist-style video games, which we deployed in an iterated learning paradigm. Game participants were limited to only two attempts (two lives) to beat each game and were allowed to write a message to a future participant who read the message before playing. Knowledge accumulated gradually across generations, allowing later generations to advance further in the games and perform more efficient actions. Multigenerational learning followed a strikingly similar trajectory to individuals learning alone with an unlimited number of lives. These results suggest that language provides a sufficient medium to express and accumulate the knowledge people acquire in these diverse tasks: the dynamics of the environment, valuable goals, dangerous risks, and strategies for success. The video game paradigm we pioneer here is thus a rich test bed for theories of cultural transmission and learning from language.

Charles Darwin's The Descent of Man was published in 1871. Ever since, it has been the foundation stone of human evolutionary studies. Richerson et al. reviewed how modern studies of human biological and cultural evolution reflect the ideas in Darwin's work. They emphasize how cooperation, social learning, and cumulative culture in the ancestors of modern humans were key to our evolution and were enhanced during the environmental upheavals of the Pleistocene. The evolutionary perspective has come to permeate not just human biology but also the social sciences, vindicating Darwin's insights. Science , aba3776, this issue p. [eaba3776][1] ### BACKGROUND Charles Darwin’s The Descent of Man , published on 24 February 1871, laid the grounds for scientific studies into human origins and evolution. We look at the advances in our understanding of these processes through the lenses of modern speciation theory. Applying this theory to specific cases requires one to identify and understand the nature of (i) the ancestor and various preexisting adaptations and traits that it possessed that allowed or simplified the speciation process, (ii) evolutionary forces responsible for major differences between the emergent species and its close relatives, and (iii) the most salient adaptations characteristic of the new species and its evolutionary history (such as genetic, morphological, behavioral, spatial, and temporal). ### ADVANCES Modern research shows that we share many developmental, physiological, morphological, cognitive, and psychological characteristics as well as about 96% of our DNA with the anthropoid apes. We now know that since our last common ancestor with the other apes 6 million to 8 million years ago, human evolution followed the path common for other species with diversification into closely related species and some subsequent hybridization between them. Since Darwin, a long series of unbridgeable gaps have been proposed between humans and other animals. They focused on tool-making, cultural learning and imitation, empathy, prosociality and cooperation, planning and foresight, episodic memory, metacognition, and theory of mind. However, new insights from neurobiology, genetics, primatology, and behavioral biology only reinforce Darwin’s view that most differences between humans and higher animals are “of degree and not of kind.” What makes us different is that our ancestors evolved greatly enhanced abilities for (and reliance on) cooperation, social learning, and cumulative culture—traits emphasized already by Darwin. Cooperation allowed for environmental risk buffering, cost reduction, and the access to new resources and benefits through the “economy of scale.” Learning and cumulative culture allowed for the accumulation and rapid spread of beneficial innovations between individuals and groups. The enhanced abilities to learn from and cooperate with others became a universal tool, removing the need to evolve specific biological organs for specific environmental challenges. These human traits likely evolved as a response to increasing high-frequency climate changes on the millennial and submillennial scales during the Pleistocene. Once the abilities for cumulative culture and extended cooperation were in place, a suite of subsequent evolutionary changes became possible and likely unavoidable. In particular, human social systems evolved to support mothers through the recruitment of males and nonreproductive females. The most distinctive feature of our species, language, appeared arguably driven by selection for simplifying cooperation. Reliance on social learning and conformity led to the emergence of new factors constraining and driving human behavior, such as morality, social norms, and social institutions. These forces often act against the immediate biological or material interests of individuals, promoting instead the interests of the society as a whole or of its powerful segments. Continuous engagement in cooperation has led to the evolution of strong coalitionary psychology, which can bring us together whenever we perceive that our identity group faces outside threats. Coalitionary psychology also has an undesirable byproduct: often negative or even hostile reaction to others who differ from us in their looks, behaviors, beliefs, caste, or class. ### OUTLOOK Our society faces challenges, including climate change; various types of inequality; economic crises; political, cultural, and religious conflicts; and pandemics. Similar challenges have repeatedly arisen and were dealt with in the past with varying success. What makes the current situation different is not only the scale of societal threats but also that modern science can provide guidance on how to respond to them. Adequately answering these challenges requires understanding humans’ social behavior and the roles of cooperation, social learning, and culture for human decision-making. Evolutionary perspective is already helping to synthesize the contributions of social sciences, including anthropology, psychology, economics, political science, and history. The impact of Descent on the social sciences and on the development and implementation of different policies by practitioners and policymakers to improve our society will only grow. ![Figure][2] Depictions of organic evolution versus cultural evolution. (Left) Organic evolution and (right) cultural evolution, as depicted in Alfred L. Kroeber’s 1923 textbook Anthropology: Cultural Patterns and Processes . Biological inheritance is rigid from parents to offspring in eukaryotes, and species mostly do not exchange genes. Culture is potentially acquired from anyone in a person’s social network, and ideas spread rather readily from culture to culture. IMAGE: N. CARY/ SCIENCE Charles Darwin’s The Descent of Man , published 150 years ago, laid the grounds for scientific studies into human origins and evolution. Three of his insights have been reinforced by modern science. The first is that we share many characteristics (genetic, developmental, physiological, morphological, cognitive, and psychological) with our closest relatives, the anthropoid apes. The second is that humans have a talent for high-level cooperation reinforced by morality and social norms. The third is that we have greatly expanded the social learning capacity that we see already in other primates. Darwin’s emphasis on the role of culture deserves special attention because during an increasingly unstable Pleistocene environment, cultural accumulation allowed changes in life history; increased cognition; and the appearance of language, social norms, and institutions. [1]: /lookup/doi/10.1126/science.aba3776 [2]: pending:yes