betaFeedback

WorksPhysics and Astronomy

Modeling the Coevolution of Human Skills and AI Adoption

Abstract

The value of relying on generative AI depends on two quantities that evolve with its use: users' unaided skill, which may decline without practice, and the originality of AI-assisted work, which may decline as more people rely on similar tools. We model this process as a coevolutionary game in which individuals choose between autonomous and AI-assisted work, with autonomous practice strengthening their skills while reliance on AI gradually erodes them. When the reinforcing effect of skill erosion outweighs the crowding cost of lost originality, the population exhibits two alternative stable states: low AI adoption with high average skill, and high AI adoption with low average skill. Because skill is individual, autonomous workers remain more skilled than AI users, making autonomous work less vulnerable to population-wide skill decline. The strength of this effect is governed by a timescale ratio: the rate at which skills adapt relative to the rate at which individuals change strategy. The same ratio determines the full distribution of skills, which we derive analytically. Populations with identical AI adoption and average skill can differ sharply: when strategies change frequently, skills remain homogeneous and collective lock-in to high AI adoption is stronger; when strategies persist, lock-in weakens but the population can polarize into highly skilled autonomous workers and AI-dependent individuals. Modeling output overlap further shows that more diverse AI outputs can make widespread adoption harder to reverse. Together, these results expose tradeoffs between the benefits of widespread AI adoption and its long-run risks, making collective lock-in and skill polarization distinct intervention targets.

The consequences of generative artificial intelligence (AI) depend not only on what people can accomplish with AI today, but also on how using it changes what they will be able to accomplish without it tomorrow. Generative AI can improve the quality and productivity of creative and knowledge-intensive work [1, 2, 3], but it may also displace the practice through which individuals develop and maintain their autonomous skills [4, 5, 6]. This raises an unresolved question at the heart of the ongoing debate about AI, i.e., whether its short-term benefits may result in the ever increasing dependence on the technology in the long-term [7]. At the same time, widespread reliance on similar AI systems may make individual outputs increasingly alike [8, 9, 10], potentially undermining the collective diversity of ideas even as individual productivity improves. Widespread AI adoption may also affect individuals unevenly, deepening skill inequalities and dividing the population into those who retain their ability to work autonomously and those who become increasingly dependent on AI [11]. With AI capabilities and patterns of use evolving rapidly, understanding how these interconnected processes might unfold requires theoretical models that can explore alternative long-term scenarios and uncover unintended collective consequences. Here, we develop such a model by studying the coevolution of AI adoption and individual skills in a population.

Existing experimental studies provide some initial insights into these concerns and tradeoffs, although most evidence concerns relatively short-term effects. In a short-story task, Doshi and Hauser found that access to generative AI improved evaluations of individual stories, particularly for writers with lower baseline creativity, while making assisted stories more similar to one another [8]. Individual gains can thus coexist with a reduction in the collective diversity of outputs. A separate set of randomized experiments by Liu et al. found that AI assistance in mathematical reasoning and reading comprehension improved assisted performance but could reduce subsequent unaided performance [5].

These findings, however, do not establish whether the short-term effects of AI assistance translate into lasting skill losses. Moreover, the consequences of AI adoption depend on how assistance is provided and whether users remain actively engaged in the underlying task [12]. In controlled decision-making experiments, AI explanations that required participants to reach their own decisions supported learning, whereas providing recommendations improved immediate decisions without comparable learning gains [13]. Related studies suggest that the design of AI assistance can either preserve or displace opportunities for independent practice [6, 14].

Understanding how these individual-level effects translate into long-term collective outcomes requires accounting for how people decide whether to use AI. Such decisions may depend on the perceived returns of autonomous and AI-assisted work, including the performance of others, and can give rise to two opposing feedback mechanisms. On the one hand, reliance on AI may displace independent practice, weakening autonomous skills and making further AI adoption increasingly attractive. On the other hand, as AI-assisted outputs become more alike, competition for originality may reduce the returns to AI adoption and discourage further use [10, 15]. The interplay between these feedbacks, together with differences in individuals’ histories of AI adoption, can shape both collective adoption patterns and the distribution of skills across the population.

Theoretical models offer a way to investigate these processes beyond the time horizons accessible to current experiments. Existing approaches have explored several relevant mechanisms, typically focusing on specific aspects of AI adoption and its consequences. At the individual level, models of cognitive offloading examine decisions to delegate work to AI when skills accumulate with practice and decay without it, showing that sufficiently substitutive AI can generate differences in skills across workers [16, 17]. Related work has identified tipping points and hysteresis in the coevolution of individual competence and tool reliance, and examined collective lock-in through a population game of outsourcing versus internalization [18]. At the population level, models have examined adoption and social welfare when widespread AI adoption affects the performance of AI systems [19], substitution and complementarity between humans and AI through cultural evolution and group selection [20], and the diffusion of large language models through social transmission, including tipping points, hysteresis, and lock-in [21]. Other work has connected AI-induced crowding of ideas to incentives for adoption [15]. Together, these studies provide insights into different parts of the problem, while leaving their interplay less well understood.

What remains missing, to the best of our knowledge, is a unified framework in which AI adoption responds to the observed returns of alternative strategies, widespread use generates competition for originality, and individual skills evolve according to each person’s history of practice. Importantly, retaining this individual dimension, rather than describing skills solely through population averages, allows us to investigate not only how AI adoption evolves collectively, but also how differences in skills emerge and develop across the population.

To that end, we develop a coevolutionary game in which individuals choose between autonomous and AI-assisted work, while their skills evolve through practice or disuse, and widespread AI adoption reduces the value of originality. Coevolutionary games couple strategies to dynamical variables that they themselves modify [22]. Previous studies have considered individual attributes such as teaching activity [23], age [24], and reputation-based partnerships [25]. In our model, the coevolving attribute is instead individual skill, whose distribution we characterize analytically.

We find that skill erosion can generate bistability and lock-in to widespread AI adoption. However, differences in individual skills can weaken this tendency, while potentially polarizing the population into highly skilled and AI-dependent individuals. The balance between these outcomes is governed by the relative timescales of skill adaptation and strategy revision. We also show that increasing the diversity of AI outputs can make widespread adoption harder to reverse. These findings highlight distinct, potentially competing objectives for interventions addressing the long-term consequences of AI adoption.

Coevolutionary game of skills and AI adoption

We consider a population of NN agents, each of whom decides whether to rely or not on AI to solve a given task, such as writing a text, generating computer code, producing a visual design, solving a quantitative problem, performing a repetitive task [26] or making a professional judgment. At time tt, each agent ii is characterized by an individual skill si(t)∈[η,1]s_i(t) \in[\eta,1], denoting its ability to perform the task without AI assistance. The parameter η∈[0,1]\eta\in[0,1] sets the lowest possible skill level.

Each agent can decide between two strategies: AI assisted work (A) or autonomous human work (H). We denote the strategy of each agent (ii) by zi(t)∈{H,A}z_i(t) \in\{H,A\}.

Agents adopting strategy HH perform the task autonomously. Their performance thus depends directly on the current individual skill, and we assign them a payoff:

πH=bhsi,(1)\pi_H = b_h s_i, \tag*{(1)}

where the parameter bh>0b_h > 0 sets the benefit associated with this choice.

Agents adopting strategy AA, instead, rely on AI assistance. We assume that AI adoption provides a baseline benefit ba>0b_a > 0, while its effectiveness decreases as adoption becomes more widespread in the population. We thus define:

πA=ba−cxα,(2)\pi_A = b_a - c x^\alpha, \tag*{(2)}

where x(t)x(t) is the fraction of agents adopting AI at time tt. The parameter c>0c > 0 controls the strength of the crowding effect, while the synergistic exponent α\alpha governs the power-law dependence on the fraction xx of AI users.

The crowding term reflects the fact that in many settings the value of an output depends not only on its quality but also on its originality. Creativity is commonly defined as the combination of novelty and usefulness [27]; scientific credit rewards priority [28]; and content that competes for limited attention loses visibility when similar contents are available [29]. An output that resembles many competing ones is therefore worth less, even when produced competently. Outputs produced with generative AI are measurably more similar to one another than those produced without it [8, 9, 10], so the crowding penalty grows with the fraction of AI users. When few people use AI, AI-assisted work remains distinctive, and its value declines as adoption spreads [15]. The cost is thus imposed by similarity to others, not by the use of AI as such. At the end of this section, we trace its origin from the overlap between human and AI-assisted outputs and discuss the effects of the synergistic exponent α>1\alpha> 1 and α<1\alpha< 1.

Agents revise their strategy through stochastic pairwise comparison [30, 31]. A focal agent ii samples another agent jj and adopts their strategy with probability

P(i←j)=11+exp⁡[−β(πj−πi)],(3)P(i \leftarrow j) = \frac{1}{1 + \exp[-\beta(\pi_j - \pi_i)]}, \tag*{(3)}

where β\beta controls the strength of payoff-based selection. This rule encodes social learning from observed performance, capturing the possible unintended consequences of AI adoption. Indeed, an agent’s payoff reflects both their individual skill and their current strategy. Thus, the imitation of strategies adopted by high-payoff-agents via Eq. (3) may disadvantage the imitating agent, as such a process is blind to individual skill [32, 33].

In addition to payoff-driven imitation, we allow for spontaneous strategy changes occurring at rate μ\mu. These changes are independent of payoff differences and represent exploration, experimentation or idiosyncratic changes in behavior. This parameter therefore introduces a persistent source of behavioral variability.

Strategy choice not only determines the immediate payoff obtained by an agent, but also affects the evolution of its ability to perform the task autonomously. We therefore let individual skill adapt over time according to the strategy currently adopted. We assume

s˙i={ε(1−si),zi=H,ε(η−si),zi=A.(4)\dot{s}_i = \begin{cases} \varepsilon(1-s_i), & z_i = H,\\ \varepsilon(\eta-s_i), & z_i = A. \end{cases} \tag*{(4)}

where ε\varepsilon controls the characteristic timescale of skill adaptation. Previous studies on skill decay under disuse imply half-lives of weeks to months for cognitive skills [18, 34], which sets the real-time scale of ε\varepsilon. Eq. (4) has the learning-forgetting structure used in recent models of AI-induced skill atrophy [16, 18, 35]. Thus, autonomous work reinforces the agent’s skill and drives sis_i to its maximal value 1, whereas the adoption of AI progressively reduces autonomous skill toward the lowest level η\eta. Since agents may switch strategy over time, their current skill depends on the history of each agent’s strategy adoption. Strategy choices {zi}\{z_i\} and skills {si}\{s_i\} therefore dynamically coevolve: the strategy currently adopted determines the evolution of an agent’s skill, where skills in turn affect the payoff associated with autonomous work and hence the relative attractiveness of the two strategies. Our microscopic agent model keeps track of the strategy and skill of every individual agent and therefore allows, in principle, for heterogeneous skills and correlations between strategy and skill. The model belongs to the class of coevolutionary games, in which agents’ strategies coevolve with additional dynamical variables describing, for instance, the state of the environment [36, 37, 38, 39] or other relevant features of the system [22, 40]. In this literature, strategies have been coupled with individual attributes, such as teaching activity [23, 41], aging [24], reputation [25].

In our model, the coevolving attribute is individual skill, which is specific to each agent and directly affects the payoff of autonomous work. This allows differences in skill levels to emerge across agents and may create correlations between strategy and skill even in an initially homogeneous population. We next derive a mean-field description of this feedback and examine how it shapes the collective dynamics.

Mean-field theory

In order to obtain analytical insights on the collective dynamics of our coevolving model of human skills and AI adoption, we derive a macroscopic description of the well-mixed population in terms of AI adoption and mean skill in the large-population and weak-selection limit (see Supplemental Material (SI)). The fraction of AI users is

x(t)=1N∑i=1N1(zi(t)=A),(5)x(t) = \frac{1}{N}\sum_{i=1}^{N}\mathbf{1}(z_i(t)=A), \tag*{(5)}

where x=0x=0 and x=1x=1 correspond, respectively, to the absence and universal adoption of AI in the population. In this mean-field limit, the dynamics can be expressed in terms of a smaller number of macroscopic variables. Due to the linearity of Eq. (4), by averaging over agents we obtain an equation for the mean skill s=N−1∑isi=(1−x)sh+xsas=N^{-1}\sum_i s_i=(1-x)s_h+xs_a, where sh=E[si∣zi=H]s_h=\mathbb{E}[s_i\mid z_i=H] and sa=E[si∣zi=A]s_a=\mathbb{E}[s_i\mid z_i=A] are the mean skills of autonomous and AI-assisted agents:

s˙=ε[1−(1−η)x−s].(6)\dot{s} = \varepsilon[1-(1-\eta)x-s]. \tag*{(6)}

Such an equation expresses the rate of change of ss as a function of ss itself and of the fraction xx of AI adopters.

In order to get an equation for the time evolution of xx, notice that the AI payoff ba−cxαb_a-cx^\alpha is the same for all agents adopting AI, whereas the payoff bhsib_hs_i of autonomous agents depends on their individual skill. Hence, every HH–AA comparison for strategy selection needs to take into account the skill of the HH agent. In the large-population and weak-selection limit, and after rescaling time by the selection intensity (see SI), the fraction of AI users obeys

x˙=x(1−x)[ba−cxα−bhsh]+m(1−2x),(7)\dot{x} = x(1-x)[b_a-cx^\alpha-b_hs_h]+m(1-2x), \tag*{(7)}

where mm is the spontaneous switching rate in rescaled units (see SI for its relation to μ\mu). Strategy selection therefore depends on shs_h rather than on ss. Selection compares AI users with the agents currently working autonomously (i.e., choosing strategy HH), that may be more skilled than the population average.

Reduced mean-field theory

We first consider a reduced description in which the mean skill of autonomous agents is approximated by the population average, sh≃ss_h\simeq s. This approximation is expected to hold when agents switch strategies much faster than their skills change, limiting the skill differences that develop between autonomous and AI-assisted agents. Under this approximation, Eq. (7) reduces to

x˙=x(1−x)[ba−cxα−bhs]+m(1−2x),(8)\dot{x}=x(1-x)\left[b_a-cx^\alpha-b_hs\right]+m(1-2x), \tag*{(8)}

which, together with Eq. (6), forms a closed system for xx and ss, that we will refer to as the reduced mean-field theory (Reduced MFT). At equilibrium, Eq. (6) gives:

s⋆=1−(1−η)x⋆.(9)s^\star=1-(1-\eta)x^\star. \tag*{(9)}

Since η<1\eta<1, higher AI adoption is therefore associated with lower mean skill. For linear crowding (α=1\alpha=1), the stationary payoff difference can be written as

Δπ⋆=πA⋆−πH⋆=ba−bh+Γx⋆,(10)\Delta\pi^\star=\pi_A^\star-\pi_H^\star=b_a-b_h+\Gamma x^\star, \tag*{(10)}

where πH⋆=bhs⋆\pi_H^\star=b_hs^\star, and

Γ=bh(1−η)−c(11)\Gamma=b_h(1-\eta)-c \tag*{(11)}

measures the balance between the skill-mediated reinforcement of AI adoption, bh(1−η)b_h(1-\eta), and the opposing effect of crowding, cc.

To clarify the role of these two terms, consider first the case without spontaneous switching, m=0m=0. As AI adoption increases, skill erosion lowers the payoff of autonomous work and therefore makes strategy AA (AI adoption) relatively more attractive. Crowding acts in the opposite direction, because a larger fraction of AI users lowers the payoff of strategy AA itself. When Γ>0\Gamma>0, the first effect dominates and AI adoption becomes self-reinforcing. For a benefit bab_a of AI adoption such that:

bhη+c<ba<bh,b_h\eta+c<b_a<b_h,

both the state (x,s)=(0,1)(x,s)=(0,1), in which all agents adopt strategy HH, and the state (1,η)(1,\eta), in which all agents adopt strategy AA (i.e. decide to use AI), are stable, and are separated by an interior saddle. When Γ<0\Gamma<0, crowding dominates instead: increasing AI adoption reduces its own relative payoff, and in the interval

bh<ba<bhη+cb_h<b_a<b_h\eta+c

the system converges to a stable interior equilibrium in which the two strategies coexist (see SI). Skill erosion alone is therefore not sufficient for bistability; its reinforcing effect must outweigh crowding.

For m>0m>0, instead, spontaneous strategy switching removes the absorbing boundary states, although it does not necessarily eliminate the two-attractor structure. In particular, there exists a range of bab_a values over which the system is bistable if and only if (see SI):

Γ>8m.(12)\Gamma>8m. \tag*{(12)}

Within this regime, two stable equilibria coexist, each with its own basin of attraction [Figure 1A]. Which equilibrium the population approaches depends on both the initial AI adoption level and the initial mean skill level: populations starting with the same adoption level can reach different long-run states, if their initial skills differ. The bistable region is bounded by two saddle-node (fold) bifurcations and gives rise to hysteresis when the baseline benefit bab_a of AI is varied slowly. As bab_a increases, the population remains on the low-adoption branch until the upper fold ba,+b_{a,+}, where the stable equilibrium collides with the saddle and disappears, forcing the system to jump to the high-adoption branch. When bab_a is subsequently decreased, the population remains on the high-adoption branch until the lower fold ba,−<ba,+b_{a,-} < b_{a,+}, where the corresponding stable equilibrium disappears and the system jumps back [Fig. 1B]. Since mean skill decreases with adoption, as per Eq. (9), it follows the corresponding loop in the opposite direction: skill drops at ba,+b_{a,+} and recovers only at ba,−b_{a,-}. Restoring an earlier value of bab_a therefore need not restore the earlier population state. Finally, increasing crowding reduces Γ\Gamma and progressively narrows the bistable region. The two folds meet at

Figure 1. Bistability and hysteresis in the reduced mean-field theory (Reduced MFT). (A) Phase portrait in the bistable regime (ba=0.6b_a = 0.6). Green solid and gray dashed curves in the (x,s)(x,s) plane are the nullclines s˙=0\dot{s}=0 and x˙=0\dot{x}=0; filled circles are the low-AI/high-skill and high-AI/low-skill equilibria; the open circle is the saddle, and the red curve is its stable manifold, which separates the basins. Blue curves are representative trajectories. (B) Equilibrium AI adoption x⋆x^\star as a function of the AI (baseline) benefit bab_a. Solid branches are stable and the dashed branch is a saddle. Arrows show the quasi-static hysteresis loop: increasing bab_a, the population remains on the low-adoption branch up to the fold ba,+≃0.925b_{a,+}\simeq0.925 and then jumps to high adoption; decreasing bab_a, it remains on the high-adoption branch down to ba,−≃0.325b_{a,-}\simeq0.325. Mean skill traces the mirror-image loop. (C) Fold points ba,±b_{a,\pm} as functions of the crowding strength cc. The shaded region is bistable and closes at the cusp where Γ=8m\Gamma=8m, c∗=bh(1−η)−8mc^\ast=b_h(1-\eta)-8m; for c>c∗c>c^\ast a single equilibrium exists for every bab_a. The vertical line marks the value of cc used in A and B. Parameters: bh=1b_h=1, η=0\eta=0, ϵ=0.05\epsilon=0.05, m=0.002m=0.002, α=1\alpha=1, c=0.25c=0.25 unless stated.

c∗=bh(1−η)−8m,c^* = b_h(1-\eta)-8m,

beyond which only a single equilibrium remains [Fig. 1C]. An analogous hysteresis behavior has been observed in a model of a single user whose reliance and competence coevolve [18]. In our model, it emerges at the population level from the competition between skill erosion and frequency-dependent crowding. Nonlinear crowding preserves this mechanism while allowing tipping between autonomy and partial AI adoption (α>1\alpha> 1), or between partial and near-complete AI adoption (α<1\alpha< 1), even when Γ≤0\Gamma\le0 (SI). Saturating crowding of the kind estimated in [15] corresponds to the latter.

Individual skills: agent-based simulations and mean-field closure

We now return to the microscopic model and simulate the coupled evolution of individual skills and strategies. This allows us to identify when the reduced MFT description applies and how emerging skill differences between strategy classes modify bistability. In our agent-based model we simulate a finite population in which each agent ii has a skill sis_i that evolves in time according to Eq. (4), and updates their strategy according to Eq. (3). Individual skills evolve continuously according to the strategy currently adopted, while strategy changes occur through imitation and spontaneous switching. The agent-based simulations therefore retain the whole distribution of individual skills and any correlations that emerge between strategies and skills. Details of the numerical simulations are provided in Methods and in the SI.

References

  1. [1]B. C. Lee and J. J. Chung, An empirical investigation of the impact of ChatGPT on creativity, Nature Human Behaviour 8, 1906 (2024).DOI
  2. [2]L. Boussioux, J. N. Lane, M. Zhang, V. Jacimovic, and K. R. Lakhani, The Crowdless Future? Generative AI and Creative Problem-Solving, Organization Science 35, 1589 (2024).DOI
  3. [3]E. Brynjolfsson, D. Li, and L. Raymond, Generative AI at Work, The Quarterly Journal of Economics 140, 889 (2025).DOI
  4. [4]S. Wu, C. G. Belem, S. Fu, M. Steyvers, and P. Smyth, How AI Assistance Affects Human Skill Development: A Study of Learning with Logic Puzzles (2026), arXiv:2608.23543 [cs.AI].arxiv.org/abs/2608.23543
  5. [5]G. Liu, B. Christian, T. Dumbalska, M. A. Bakker, and R. Dubey, AI Assistance Reduces Persistence and Hurts Independent Performance (2026), arXiv:2604.04721 [cs.AI].arxiv.org/abs/2604.04721
  6. [6]H. Bastani, O. Bastani, A. Sungu, H. Ge, Ö. Kabakcı, and R. Mariman, Generative AI without guardrails can harm learning: Evidence from high school mathematics, Proceedings of the National Academy of Sciences 122, e2422633122 (2025).DOI
  7. [7]T. N. Cash, M. O. Kelly, B. N. Macnamara, and E. F. Risko, Is AI making us stupid?, Trends in Cognitive Sciences 30, 673 (2026).DOI
  8. [8]A. R. Doshi and O. P. Hauser, Generative AI enhances individual creativity but reduces the collective diversity of novel content, Science Advances 10, eadn5290 (2024).DOI
  9. [9]L. Meincke, G. Nave, and C. Terwiesch, ChatGPT decreases idea diversity in brainstorming, Nature Human Behaviour 9, 1107 (2025).DOI
  10. [10]K. Moon, A. E. Green, and K. Kushlev, Homogenizing effect of large language models (LLMs) on creative diversity: An empirical comparison of human and ChatGPT writing, Computers in Human Behavior: Artificial Humans 6, 100207 (2025).
  11. [11]S. Valcea, M. R. Hamdani, and S. Wang, Exploring the Impact of ChatGPT on Business School Education: Prospects, Boundaries, and Paradoxes, Journal of Management Education 48, 915 (2024).DOI
  12. [12]Møller, A. G., Bassignana, E., Pierri, F., & Aiello, L. M. (2026). Overreliance on AI in Information-seeking from Video Content.arXiv
  13. [13]K. Z. Gajos and L. Mamykina, Do People Engage Cognitively with AI? Impact of AI Assistance on Incidental Learning, in Proceedings of the 27th International Conference on Intelligent User Interfaces, IUI ’22 (Association for Computing Machinery, New York, NY, USA, 2022) pp. 794–806.DOI
  14. [14]B. Lira, T. Rogers, D. G. Goldstein, L. Ungar, and A. L. Duckworth, Coach not crutch: Evidence that AI can improve writing skill despite reducing effort (2026), arXiv:2502.02880 [cs.HC].arxiv.org/abs/2502.02880
  15. [15]N. S. Azad and R. A. Baten, Ex Ante Evaluation of AI-induced Idea Diversity Collapse (2026), arXiv:2605.06540 [cs.AI].arxiv.org/abs/2605.06540
  16. [16]M. Caosun and S. Aral, The Augmentation Trap: AI Productivity and the Cost of Cognitive Offloading (2026), arXiv:2604.03501 [cs.HC].arxiv.org/abs/2604.03501
  17. [17]Afrouzi, H., Blanco, A., Drenik, A., & Hurst, E. (2026). Automation, Learning, and Career Dynamics.
  18. [18]D. C. Krakauer, Competitive and Complementary Tools (2026), arXiv:2607.18460 [q-bio.NC].arxiv.org/abs/2607.18460
  19. [19]F. Baumann, E. Akçay, and J. B. Plotkin, Generative artificial intelligence reduces social welfare through model collapse (2026), arXiv:2604.21853 [physics.soc-ph].arxiv.org/abs/2604.21853
  20. [20]Zhong, Q., Shon, T. F., Schoenmann, J., Garcia, J., & Rahwan, R. (2026). Group Selection as a Safeguard Against AI Substitution.arXiv
  21. [21]R. Solé, G. Ruffini, F. Castaldo, M. Tuccio, L. F. Seoane, M. d. Domenico, S. F. Elena, D. C. Krakauer, and M. Levin, Large-Language Models as a Cognitive Virus (2026), arXiv:2609.03344 [physics.soc-ph].arxiv.org/abs/2609.03344
  22. [22]M. Perc and A. Szolnoki, Coevolutionary games—A mini review, Biosystems 99, 109 (2010).DOI
  23. [23]A. Szolnoki and M. Perc, Coevolution of teaching activity promotes cooperation, New Journal of Physics 10, 043036 (2008).DOI
  24. [24]A. Szolnoki, M. Perc, G. Szabó, and H.-U. Stark, Impact of aging on the evolution of cooperation in the spatial prisoner’s dilemma game, Physical Review E 80, 021901 (2009).pubmed.ncbi.nlm.nih.gov/19792145
  25. [25]F. Fu, C. Hauert, M. A. Nowak, and L. Wang, Reputation-based partner choice promotes cooperation in social networks, Physical Review E 78, 026117 (2008).pubmed.ncbi.nlm.nih.gov/18850907
  26. [26]Srivastava, R., Abhinandan, A., Verma, A., & Gupta, Y. (2026). Agentic Browsers: A New Paradigm for Repetitive Web Task Automation, in 2026 International Conference on Sustainable Engineering and Technology Innovations (ICSETI). (787–791).
  27. [27]M. A. Runco and G. J. Jaeger, The Standard Definition of Creativity, Creativity Research Journal 24, 92 (2012).DOI
  28. [28]R. K. Merton, Priorities in Scientific Discovery: A Chapter in the Sociology of Science, American Sociological Review 22, 635 (1957).DOI
  29. [29]L. Wang, A. Flammini, A. Vespignani, and F. Menczer, Competition among memes in a world with limited attention, Scientific Reports 2, 335 (2012).DOI
  30. [30]A. Traulsen, M. A. Nowak, and J. M. Pacheco, Stochastic dynamics of invasion and fixation, Physical Review E 74, 011909 (2006).pubmed.ncbi.nlm.nih.gov/16907129
  31. [31]A. Traulsen, J. M. Pacheco, and M. A. Nowak, Pairwise comparison and selection temperature in evolutionary game dynamics, Journal of Theoretical Biology 246, 522 (2007).pubmed.ncbi.nlm.nih.gov/17292423
  32. [32]G. Livan, Don’t follow the leader: how ranking performance reduces meritocracy, Royal Society Open Science 6, 191255 (2019).DOI
  33. [33]De Domenico, F., Caccioli, F., Livan, G., Montagna, G., & Nicosini, O. (2024). Imitation versus serendipity in ranking dynamics. Royal Society Open Science, 11, 240177.
  34. [34]W. Arthur Jr., W. Bennett Jr., P. L. Stanush, and T. L. McNelly, Factors That Influence Skill Decay and Retention: A Quantitative Review and Analysis, Human Performance 11, 57 (1998).DOI
  35. [35]V. R. R. Ganuthula and M. K. Singh, The Paradox of Augmentation: A Theoretical Model of AI-Induced Skill Atrophy, Human Behavior and Emerging Technologies 2026, 8303770 (2026).DOI
  36. [36]A. R. Tilman, J. B. Plotkin, and E. Akçay, Evolutionary games with environmental feedbacks, Nature Communications 11, 915 (2020).DOI
  37. [37]C. Hauert, C. Saade, and A. McAvoy, Asymmetric evolutionary games with environmental feedback, Journal of Theoretical Biology 462, 347 (2019).DOI
  38. [38]J. S. Weitz, C. Eksin, K. Paarporn, S. P. Brown, and W. C. Ratcliff, An oscillating tragedy of the commons in replicator dynamics with game-environment feedback, Proceedings of the National Academy of Sciences 113, E7518 (2016).pubmed.ncbi.nlm.nih.gov/27830651
  39. [39]Q. Su, A. McAvoy, L. Wang, and M. A. Nowak, Evolutionary dynamics with game transitions, Proceedings of the National Academy of Sciences 116, 25398 (2019).DOI
  40. [40]J. M. Pacheco, A. Traulsen, and M. A. Nowak, Coevolution of Strategy and Structure in Complex Networks with Dynamical Linking, Physical Review Letters 97, 258103 (2006).pubmed.ncbi.nlm.nih.gov/17280398
  41. [41]A. Szolnoki and G. Szabó, Cooperation enhanced by in-homogeneous activity of teaching for evolutionary Prisoner’s Dilemma games, Europhysics Letters 77, 30004 (2007).DOI
  42. [42]J. Peccoud and B. Ycart, Markovian Modeling of Gene-Product Synthesis, Theoretical Population Biology 48, 222 (1995).DOI
  43. [43]Horsthemke, W., & Lefever, R. (1984). Noise-Induced Transitions. In Springer Series in Synergetics (Vol. 15).
  44. [44]J. M. Sancho, Stochastic processes driven by dichotomous Markov noise: Some exact dynamical results, Journal of Mathematical Physics 25, 354 (1984).DOI
  45. [45]Ashkinaze, J., Mendelsohn, J., Qiwei, L., Budak, C., & Gilbert, E. (2025). How AI Ideas Affect the Creativity, Diversity, and Evolution of Human Ideas: Evidence From a Large, Dynamic Experiment. Proceedings of the ACM Collective Intelligence Conference, CI ’25, 198–213.
  46. [46]E. H. Simpson, Measurement of Diversity, Nature 163, 688 (1949).DOI
  47. [47]Z. Contractor and G. Reyes, Generative AI in Higher Education: Evidence from an Elite College (2026), arXiv:2508.00717 [econ.GN].arxiv.org/abs/2508.00717
  48. [48]G. Charness and D. Grieco, Creativity and AI, The Economic Journal , ueag015 (2026).DOI
  49. [49]I. Shumailov, Z. Shumaylov, Y. Zhao, N. Papernot, R. Anderson, and Y. Gal, AI models collapse when trained on recursively generated data, Nature 631, 755 (2024).DOI
  50. [50]M. Gerstgrasser, R. Schaeffer, A. Dey, R. Rafailov, H. Sleight, J. Hughes, T. Korbak, R. Agrawal, D. Pai, A. Gromov, D. A. Roberts, D. Yang, D. L. Donoho, and S. Koyejo, Is Model Collapse Inevitable? Breaking the Curse of Recursion by Accumulating Real and Synthetic Data (2024), arXiv:2404.01413 [cs.LG].arXiv
  51. [51]F. Tria, V. Loreto, V. D. P. Servedio, and S. H. Strogatz, The dynamics of correlated novelties, Scientific Reports 4, 5890 (2014).DOI
  52. [52]V. Loreto, V. D. P. Servedio, S. H. Strogatz, and F. Tria, Dynamics on Expanding Spaces: Modeling the Emergence of Novelties, in Creativity and Universality in Language, edited by M. Degli Esposti, E. G. Altmann, and F. Pachet (Springer International Publishing, Cham, 2016) pp. 59–83.DOI
  53. [53]I. Iacopini, S. Milojević, and V. Latora, Network Dynamics of Innovation Processes, Physical Review Letters 120, 048301 (2018).DOI
  54. [54]G. Di Bona, A. Bellina, G. De Marzo, A. Petralia, I. Iacopini, and V. Latora, The dynamics of higher-order novelties, Nature Communications 16, 393 (2025).DOI
  55. [55]J. Park, M. Kim, and S. Han, The enrichment paradox: critical capability thresholds and irreversible dependency in human-AI symbiosis (2026), arXiv:2603.24391 [cs.CY].arXiv
  56. [56]I. Bena, Dichotomous Markov noise: Exact results for out-of-equilibrium systems. A review, International Journal of Modern Physics B 20, 2825 (2006), arXiv:cond-mat/0606116.DOI
  57. [57]Strogatz, S. H. (2014). Nonlinear Dynamics and Chaos: With Applications to Physics, Biology, Chemistry, and Engineering.

Paper details

vlnc:W_8h35g89p
Version
v1 · latest
Published
Added
License
CC BY 4.0
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents