Decision-making on numerous aspects of our daily lives is being outsourced to machine-learning (ML) algorithms and artificial intelligence (AI), motivated by speed and efficiency in the decision process. ML approaches—one of the typologies of algorithms underpinning artificial intelligence—are typically developed as black boxes. The implication is that ML code scripts are rarely scrutinised; interpretability is usually sacrificed in favour of usability and effectiveness. Room for improvement in practices associated with programme development have also been flagged along other dimensions, includinginter aliafairness, accuracy, accountability, and transparency. In this contribution, the production of guidelines and dedicated documents around these themes is discussed. The following applications ofAI-drivendecision-making are outlined: (a) risk assessment in the criminal justice system, and (b) autonomous vehicles, highlighting points of friction across ethical principles. Possible ways forward towards the implementation of governance on AI are finally examined.

The reasons for and against composite indicators are briefly reviewed, as well as the available theories for their construction. After noting the strong normative dimension of these measures—which ultimately aim to ‘tell a story’, e.g. to promote the social discovery of a particular phenomenon, we inquire whether a less partisan use of a composite indicator can be proposed by allowing more latitude in the framing of its construction. We thus explore whether a composite indicator can be built to tell ‘more than one story’ and test this in practical contexts. These include measures used in convergence analysis in the field of cohesion policies and a recent case involving the World Bank’s Doing Business Index. Our experiments are built to imagine different constituencies and stakeholders who agree on the use of evidence and of statistical information while differing on the interpretation of what is relevant and vital.

Mathematical models are getting increasingly detailed to better predict phenomena or gain more accurate insights into the dynamics of a system of interest, even when there are no validation or training data available. Here, we show through ANOVA and statistical theory that this practice promotes fuzzier estimates because it generally increases the model’s effective dimensions, i.e., the number of influential parameters and the weight of high-order interactions. By tracking the evolution of the effective dimensions and the output uncertainty at each model upgrade stage, modelers can better ponder whether the addition of detail truly matches the model’s purpose and the quality of the data fed into it.

Predictions of global irrigated areas are widely used to guide strategies that aim to secure environmental welfare and manage climate change. Here we show that these predictions, which range between 240 and 450 million hectares (Mha), underestimate the potential extension of irrigation by ignoring basic parametric and model uncertainties. We found that the probability distribution of global irrigated areas in 2050 spans almost half an order of magnitude (∼300–800 Mha, P_2.5,P_97.5), with the right tail pushing values to up to ∼1,800 Mha. This uncertainty is mostly irreducible as it is largely caused by either population‐related parameters or the assumptions behind the model design. Model end‐users and policy makers should acknowledge that irrigated areas are likely to grow much more than previously thought in order to avoid underestimating potential environmental costs.