Although artificial intelligence (AI) has become increasingly smart, its wisdom has not kept pace. In this opinion article, we examine what is known about human wisdom and sketch a vision of its AI counterpart. We introduce human wisdom as strategies for solving intractable problems-those outside the scope of analytic techniques-including both 'object-level' strategies, such as heuristics (for managing problems), and 'metacognitive' strategies, such as intellectual humility, perspective-taking, or context adaptability (for managing object-level task fit). We argue that AI systems particularly struggle with this type of metacognition. Wise metacognition would lead to AI that is more robust to novel environments, explainable to users, cooperative with others, and safer by risking fewer misaligned goals with human users. We discuss how wise AI might be benchmarked, trained, and implemented.

The theory of thermal macroeconomics (TM) analyses economic phenomena within the mathematical framework of classical thermodynamics, using a set of axioms that apply to the purely macroscopic aspects of an economy [CM]. The theory shows that the possible macro-behaviours are governed by an entropy function. In simple idealised cases, the entropy function can be calculated from the rules governing the interactions of individual agents. But where this is not possible, TM predicts that the entropy can nonetheless be measured empirically through an economic analogue of calorimetry in physics. We show using computer simulations the in-principle feasibility of this approach: an entropy function can successfully be measured for a range of simulated economies that we tested. In cases where entropy can be calculated analytically from microfoundational assumptions, the measured entropy agrees well. In more complex cases, where microfoundational analysis is infeasible, our method of measuring entropy still applies and is validated by demonstrations that entropy is a state function of an economic system, i.e., exhibits path independence. This appears to hold even for some systems to which we don't have a proof that the Axioms of TM apply. Furthermore, in all cases tested, entropy is concave, as predicted by TM. As shown in [CM], once the entropy function is established for a simulated exchange economy, it is possible to derive prices, the value of money and various other quantities, and make predictions about the effects of putting two or more economies in contact.

The commentaries collectively provide a thoughtful and constructive analysis of resource rational contractualism (RRC), helping both to push the theory forward and clarifying its limitations. We use these as a springboard to revise and improve upon the ideas presented in the target article. Our replies are organized in three main themes: possible extensions, challenges, and wider applications.

Human societies rely on shared conventions that allow people to coordinate efficiently, from the conversational etiquette and the “rules of the road” to workplace roles and the handshake economy. To be effective, the conventions must be reliable, and yet also adaptable: when circumstances change, people must decide whether to follow established precedents or shift to new ways of coordinating. Most cognitive and computational models explain either how coordination equilibria emerge and persist through learning from past experience, or how agents generate novel strategies to maximize anticipated mutual benefit. How people arbitrate between these two modes of coordination remains poorly understood. Here we introduce a computational model and experimental paradigm to study this arbitration. In two preregistered experiments (n = 510; 30,420 choices), participants repeatedly solved coordination problems in which established precedents could be followed or abandoned as their efficiency declined. Our results quantify the conditions under which individuals and pairs transition from entrenched coordination equilibria to new shared solutions, revealing a dynamic trade-off between the reliability of precedent and the potential gains from innovation. These findings provide a unified account of how coordination remains both stable and adaptable, offering new insights into the evolution of conventions, norms, and institutions.

Human coordination depends on two complementary mechanisms: forward-looking strategies that enable flexible adaptation to new circumstances, and backward-looking mechanisms that rely on precedent, convention, and rule-following. Most cognitive and computational models of coordination emphasize one mechanism or the other—either explaining how equilibria emerge and persist when agents adapt their behavior based on past experience, or how agents creatively generate novel solutions and strategies to achieve anticipated mutual benefit in the challenges of the moment—but not how the two interact. Here we introduce a cognitive model and experimental paradigm to capture the dynamics of both processes and, crucially, the arbitration between them. In two preregistered experiments (n = 510; 30,420 choices), participants repeatedly solve coordination problems that can be addressed either by generalizing past solutions or by adopting novel ones when precedent becomes inefficient. This design allows us to examine the conditions under which individuals or dyads decide to abandon entrenched equilibria and transition to novel coordination solutions by arbitrating between mutual benefit and precedent. By formally modeling both forward- and backward-looking mechanisms, and the process of arbitration between them, we provide a unified framework for understanding how human coordination can be both stable and adaptable—a property that underlies everyday cooperative behavior, social norms, and institutional evolution.

In this paper, we test predictions of a new theory of macroeconomics, called "thermal macroeconomics." The theory aims to apply the mathematical structure of classical thermodynamics, including analogues of temperature and entropy, to predict aspects of the aggregate behaviour of populations of economic agents without analyzing their detailed interactions. We test the theory by comparing its predictions with the behaviour of a variety of simulated micro-economies in which goods and money can be exchanged between agents, confirming the predictions of the theory. The paper serves also to illustrate and make more tangible the predictions of thermal macroeconomics.

Does the utility of an outcome influence people’s assessment of risk and uncertainty? Growing evidence suggests that people often rely on mental simulations to evaluate probability and risky events. However, prior experimental findings offer conflicting predictions about how utility biases this mental sampling process. Across four experiments (total N=206, with Experiment 4 pre-registered), we investigated the influence of utility using a random generation paradigm. These responses were then compared to probability judgments and predictions. While we identified individual differences, the majority of participants exhibited neutrality, with no systematic impact of utility on their sampling distributions. Nevertheless, biases emerged under specific conditions, including a preference for smaller or more probable outcomes as the starting point of simulations and optimism in single-response predictions. Additionally, we found evidence suggesting that probability judgments, predictions, and random generation tasks may rely on a shared underlying mental process. Our findings suggest that models of judgment and decision-making should account for individual differences in utility influences, particularly distinguishing between unbiased sampling and optimistic sampling—the selective over-representation of high-utility outcomes.

People frequently face decisions that require making inferences about withheld information. The advent of large language models coupled with conversational technology, e.g., Alexa, Siri, Cortana, and the Google Assistant, is changing the mode in which people make these inferences. We demonstrate that conversational modes of information provision, relative to traditional digital media, result in more critical responses to withheld information, including: (1) a reduction in evaluations of a product or service for which information is withheld and (2) an increased likelihood of recalling that information was withheld. These effects are robust across multiple conversational modes: a recorded phone conversation, an unfolding chat conversation, and a conversation script. We provide further evidence that these effects hold for conversations with the Google Assistant, a prominent conversational technology. The experimental results point to participants’ intuitions about why the information was withheld as the driver of the effect.

Repeated forecasts of changing values are a key aspect of many everyday tasks, from predicting the weather to financial markets. A particularly simple and informative instance of such fluctuating values are random walks: sequences in which each point is a random movement from only its preceding value, unaffected by any previous points. Moreover, random walks often yield basic rational forecasting solutions in which predictions of new values should repeat the most recent value, and hence replicate the properties of the original series. In previous experiments, however, we have found that human forecasters do not adhere to this standard, showing systematic deviations from the properties of a random walk such as excessive volatility and extreme movements between subsequent predictions. We suggest that such deviations reflect general statistical signatures of human cognition displayed across multiple tasks, offering a window into underlying cognitive mechanisms. Using these deviations as new criteria, we here explore several cognitive models of forecasting drawn from various approaches developed in the existing literature, including Bayesian, error-based learning, autoregressive and sampling mechanisms. These models are contrasted with human data from two experiments to determine which best accounts for the particular statistical features displayed by participants. We find support for sampling models in both aggregate and individual fits, suggesting that these variations are attributable to the use of inherently stochastic prediction systems. We thus argue that variability in predictions is primarily driven by computational noise within the decision making process, rather than "late" noise at the output stage.

It is widely agreed upon that morality guides people with conflicting interests towards agreements of mutual benefit. We therefore might expect numerous proposals for organizing human moral cognition around the logic of bargaining, negotiation, and agreement. Yet, while “contractualist” ideas play an important role in moral philosophy, they are starkly underrepresented in the field of moral psychology. From a contractualist perspective, ideal moral judgments are those that would be agreed to by rational bargaining agents—an idea with wide-spread support in philosophy, psychology, economics, biology, and cultural evolution. As a practical matter, however, investing time and effort in negotiating every interpersonal interaction is unfeasible. Instead, we propose, people use abstractions and heuristics to efficiently identify mutually beneficial arrangements. We argue that many well-studied elements of our moral minds, such as reasoning about others’ utilities (“consequentialist” reasoning) or evaluating intrinsic ethical properties of certain actions (“deontological” reasoning), can be naturally understood as resource-rational approximations of a contractualist ideal. Moreover, this view explains the flexibility of our moral minds—how our moral rules and standards get created, updated and overridden and how we deal with novel cases we have never seen before. Thus, the apparently fragmentary nature of our moral psychology—commonly described in terms of systems in conflict—can be largely unified around the principle of finding mutually beneficial agreements under resource constraint. Our resulting “triple theory” of moral cognition naturally integrates contractualist, consequentialist and deontological concerns.

In many tasks, human behavior is far noisier than is optimal. Yet when asked to behave randomly, people are typically too predictable. We argue that these apparently contrasting observations have the same origin: the operation of a general-purpose local sampling algorithm for probabilistic inference. This account makes distinctive predictions regarding random sequence generation, not predicted by previous accounts -- which suggests that randomness is produced by inhibition of habitual behavior, striving for unpredictability. We verify these predictions in two experiments: people show the same deviations from randomness when randomly generating from non-uniform or recently-learned distributions. In addition, our data show a novel signature behavior, that people's sequences have too few changes of trajectory, which argues against the specific local sampling algorithms that have been proposed in past work with other tasks. Using computational modeling, we show that local sampling where direction is maintained across trials best explains our data, which suggests it may be used in other tasks too. While local sampling has previously explained why people are unpredictable in standard cognitive tasks, here it also explains why human random sequences are not unpredictable enough.

People solve a myriad of coordination problems without explicit communication every day. A recent theoretical account, virtual bargaining, proposes that, to coordinate, we often simulate a negotiation process, and act according to what we would be most likely to agree to do if we were to bargain. But very often several equivalent tacit agreements — or virtual bargains — are available, which poses the challenge of figuring out which one to follow. Here we take inspiration from virtual bargaining to develop a cognitive modeling framework for dynamic coordination problems. We assume that players recognize their common goal, identify one or more possible tacit agreements based on situational features, observe the history of their partner’s choices to infer the most likely tacit agreement, and play their role in the joint plan. We test this approach in two experiments (n = 125 and n = 133) based on a dynamic coordination game designed to elicit agreement-based behavior. We fit our model at the individual level and compare its performance against alternative models. Across four different conditions, our model performs best among the set of models considered. Behavioral results are also consistent with players sustaining coordination and cooperation in the task by converging on tacitly agreed strategies or “virtual bargains”.

Psychological variability (i.e., "noise") displays interesting structure which is hidden by the common practice of averaging over trials. Interesting noise structure, termed 'stylized facts', is observed in financial markets (i.e., behaviors from many thousands of traders). Here we investigate the parallels between psychological and financial time series. In a series of three experiments (total N = 202), we successively simplified a market-based price prediction task by first removing external information, and then removing any interaction between participants. Finally, we removed any resemblance to an asset market by asking individual participants to simply reproduce temporal intervals. All three experiments reproduced the main stylized facts found in financial markets, and the robustness of the results suggests that a common cognitive-level mechanism can produce them. We identify one potential model based on mental sampling algorithms, showing how this general-purpose model might account for behavior across these very different tasks.