Eladio Montero-Porras

Recommender systems are not neutral tools; they are AI systems that actively shape human behaviour while being shaped by it in return. When platforms prioritise short-term revenue over content quality, these human-algorithm feedback loops can push ecosystems into self-reinforcing decline. This paper presents a game-theoretic model of co-adaptation between a platform deploying an algorithmic recommender and a population of creators responding to its incentives. The model yields two equilibria: a cooperative outcome with sustained high-quality content, and a defective outcome in which both parties prioritise short-term gains. The transition between them can be irreversible: a temporary relaxation of quality standards may shift the system beyond a tipping point from which restoring the original policy no longer recovers the prior equilibrium. The location of this tipping point depends on the initial health of the ecosystem, implying that early intervention is structurally more effective than delayed correction. The model is intentionally minimal, providing a baseline for future extensions while highlighting how human-AI system design can generate persistent societal effects.

Managing shared resources requires balancing personal profit and sustainability. This paper reports on a behavioural experiment testing how extraction defaults—either pro-social or exploitative—impact resource extraction in a common pool resource dilemma (CPRD). We find that an exploitative default increases average extraction compared to a control without a default, while a pro-social default temporarily reduces extraction. The effects of both defaults are temporary, and extraction levels converge to those in the control group, with the pro-social default fading faster. Notably, the influence of defaults depended on individual inclinations, with cooperative individuals extracting more under an exploitative default, and selfish individuals less under a pro-social default. Our findings suggest that while defaults can promote short-term sustainability, their long-term effects are limited, and their effectiveness depends on individual traits.

Extracting from shared resources requires making choices to balance personal profit and sustainability. We present the results of a behavioural experiment wherein we manipulate the default extraction from a finite resource. Participants were exposed to two treatments -- pro-social or self-serving extraction defaults -- and a control without defaults. We examined the persistence of these nudges by removing the default after five rounds. Results reveal that a self-serving default increased the average extraction while present, whereas a pro-social default only decreased extraction for the first two rounds. Notably, the influence of defaults depended on individual inclinations, with cooperative individuals extracting more under a self-serving default, and selfish individuals less under a pro-social default. After the removal of the default, we observed no significant differences with the control treatment. Our research highlights the potential of defaults as cost-effective tools for promoting sustainability, while also advocating for a careful use to avoid adverse effects.

People have different preferences for what they allocate for themselves and what they allocate to others in social dilemmas. These differences result from contextual reasons, intrinsic values, and social expectations. What is still an area of debate is whether these differences can be estimated from differences in each individual’s deliberation process. In this work, we analyse the participants’ reaction times in three different experiments of the Iterated Prisoner’s Dilemma with the Drift Diffusion Model, which links response times to the perceived difficulty of the decision task, the rate of accumulation of information (deliberation), and the intuitive attitudes towards the choices. The correlation between these results and the attitude of the participants towards the allocation of resources is then determined. We observe that individuals who allocated resources equally are correlated with more deliberation than highly cooperative or highly defective participants, who accumulate evidence more quickly to reach a decision. Also, the evidence collection is faster in fixed neighbour settings than in shuffled ones. Consequently, fast decisions do not distinguish cooperators from defectors in these experiments, but appear to separate those that are more reactive to the behaviour of others from those that act categorically.

While many theoretical studies have revealed the strategies that could lead to and maintain cooperation in the Iterated Prisoner’s dilemma, less is known about what human participants actually do in this game and how strategies change when being confronted with anonymous partners in each round. Previous attempts used short experiments, made different assumptions of possible strategies, and led to very different conclusions. We present here two long treatments that differ in the partner matching strategy used, i.e. fixed or shuffled partners. Here we use unsupervised methods to cluster the players based on their actions and then Hidden Markov Model to infer what the memory-one strategies are in each cluster. Analysis of the inferred strategies reveals that fixed partner interaction leads to behavioral self-organization. Shuffled partners generate subgroups of memory-one strategies that remain entangled, apparently blocking the self-selection process that leads to fully cooperating participants in the fixed partner treatment. Analyzing the latter in more detail shows that AllC, AllD, TFT- and WSLS-like behavior can be observed. This study also reveals that long treatments are needed as experiments with less than 25 rounds capture mostly the learning phase participants go through in these kinds of experiments.

Our understanding of how a city works is changing as more and more activities and interactions of citizens with and within the city are being tracked. It is estimated that by 2020, 1.7MB of data will be created every second for each person [1]. These data are created by different actors, including businesses that collect information to fuel their services. Cities can enrich the understanding of their citizen dynamics and improve data-driven decision-making using these data. However, cities do not have the resources or skills (yet) to handle and analyze this new type of information. In this paper, we propose a methodology for cities to map and identify different sources of existing data, in order to give them a new meaning in the urban planning context. Also, we present our empirical experience with the methodology in solving issues at the city management level. Results show that cities can not only gather new insights by exploring existing datasets from businesses but also give them a new purpose to support and evaluate decision-making on the urban space and its development.