David Watson

Jakob Mokander, Prathm Juneja, David Watson et al.

On the whole, the U.S. Algorithmic Accountability Act of 2022 (US AAA) is a pragmatic approach to balancing the benefits and risks of automated decision systems. Yet there is still room for improvement. This commentary highlights how the US AAA can both inform and learn from the European Artificial Intelligence Act (EU AIA).

Jakob Mokander, Margi Sheth, David Watson et al.

Organisations that design and deploy artificial intelligence (AI) systems increasingly commit themselves to high-level, ethical principles. However, there still exists a gap between principles and practices in AI ethics. One major obstacle organisations face when attempting to operationalise AI Ethics is the lack of a well-defined material scope. Put differently, the question to which systems and processes AI ethics principles ought to apply remains unanswered. Of course, there exists no universally accepted definition of AI, and different systems pose different ethical challenges. Nevertheless, pragmatic problem-solving demands that things should be sorted so that their grouping will promote successful actions for some specific end. In this article, we review and compare previous attempts to classify AI systems for the purpose of implementing AI governance in practice. We find that attempts to classify AI systems found in previous literature use one of three mental model. The Switch, i.e., a binary approach according to which systems either are or are not considered AI systems depending on their characteristics. The Ladder, i.e., a risk-based approach that classifies systems according to the ethical risks they pose. And the Matrix, i.e., a multi-dimensional classification of systems that take various aspects into account, such as context, data input, and decision-model. Each of these models for classifying AI systems comes with its own set of strengths and weaknesses. By conceptualising different ways of classifying AI systems into simple mental models, we hope to provide organisations that design, deploy, or regulate AI systems with the conceptual tools needed to operationalise AI governance in practice.

Edoardo Pona, Milad Kazemi, Mehran Hosseini et al.

Monitoring LLM safety at scale requires balancing cost and accuracy: a cheap latent-space probe can screen every input, but hard cases should be escalated to a more expensive expert. Existing cascades delegate based on probe uncertainty, but uncertainty is a poor proxy for delegation benefit, as it ignores whether the expert would actually correct the error. To address this problem, we introduce Calibrate-Then-Delegate (CTD), a model-cascade approach that provides probabilistic guarantees on the computation cost while enabling instance-level (streaming) decisions. CTD builds on a novel delegation value (DV) probe, a lightweight model operating on the same internal representations as the safety probe that directly predicts the benefit of escalation. To enforce budget constraints, CTD calibrates a threshold on the DV signal using held-out data via multiple hypothesis testing, yielding finite-sample guarantees on the delegation rate. Evaluated on four safety datasets, CTD consistently outperforms uncertainty-based delegation at every budget level, avoids harmful over-delegation, and adapts budget allocation to input difficulty without requiring group labels.

Luciano Floridi, Jessica Morley, Claudio Novelli et al.

This article looks at how reasoning works in current Large Language Models (LLMs) that function using the token-completion method. It examines their stochastic nature and their similarity to human abductive reasoning. The argument is that these LLMs create text based on learned patterns rather than performing actual abductive reasoning. When their output seems abductive, this is largely because they are trained on human-generated texts that include reasoning structures. Examples are used to show how LLMs can produce plausible ideas, mimic commonsense reasoning, and give explanatory answers without being grounded in truth, semantics, verification, or understanding, and without performing any real abductive reasoning. This dual nature, where the models have a stochastic base but appear abductive in use, has important consequences for how LLMs are evaluated and applied. They can assist with generating ideas and supporting human thinking, but their outputs must be critically assessed because they cannot identify truth or verify their explanations. The article concludes by addressing five objections to these points, noting some limitations in the analysis, and offering an overall evaluation.

Michelle Seng Ah Lee, Kirtan Padh, David Watson et al.

Fair machine learning (ML) methods help identify and mitigate the risk that algorithms encode or automate social injustices. Algorithmic approaches alone cannot resolve structural inequalities, but they can support socio-technical decision systems by surfacing discriminatory biases, clarifying trade-offs, and enabling governance. Although fairness is well studied in supervised learning, many real ML applications are online and sequential, with prior decisions informing future ones. Each decision is taken under uncertainty due to unobserved counterfactuals and finite samples, with dire consequences for under-represented groups, systematically under-observed due to historical exclusion and selective feedback. A bank cannot know whether a denied loan would have been repaid, and may have less data on marginalized populations. This paper introduces a taxonomy of uncertainty in sequential decision-making -- model, feedback, and prediction uncertainty -- providing shared vocabulary for assessing systems where uncertainty is unevenly distributed across groups. We formalize model and feedback uncertainty via counterfactual logic and reinforcement learning, and illustrate harms to decision makers (unrealized gains/losses) and subjects (compounding exclusion, reduced access) of policies that ignore the unobserved space. Algorithmic examples show it is possible to reduce outcome variance for disadvantaged groups while preserving institutional objectives (e.g. expected utility). Experiments on data simulated with varying bias show how unequal uncertainty and selective feedback produce disparities, and how uncertainty-aware exploration alters fairness metrics. The framework equips practitioners to diagnose, audit, and govern fairness risks. Where uncertainty drives unfairness rather than incidental noise, accounting for it is essential to fair and effective decision-making.

Edoardo Pona, Milad Kazemi, Yali Du et al.

Counterfactual inference is a powerful tool for analysing and evaluating autonomous agents, but its application to language model (LM) agents remains challenging. Existing work on counterfactuals in LMs has primarily focused on token-level counterfactuals, which are often inadequate for LM agents due to their open-ended action spaces. Unlike traditional agents with fixed, clearly defined action spaces, the actions of LM agents are often implicit in the strings they output, making their action spaces difficult to define and interpret. Furthermore, the meanings of individual tokens can shift depending on the context, adding complexity to token-level reasoning and sometimes leading to biased or meaningless counterfactuals. We introduce \emph{Abstract Counterfactuals}, a framework that emphasises high-level characteristics of actions and interactions within an environment, enabling counterfactual reasoning tailored to user-relevant features. Our experiments demonstrate that the approach produces consistent and meaningful counterfactuals while minimising the undesired side effects of token-level methods. We conduct experiments on text-based games and counterfactual text generation, while considering both token-level and latent-space interventions.

Stephen Asiedu, David Watson

Causal discovery in multi-omic datasets is crucial for understanding the bigger picture of gene regulatory mechanisms, but remains challenging due to high dimensionality, differentiation of direct from indirect relationships, and hidden confounders. We introduce GENESIS (GEne Network inference from Expression SIgnals and SNPs), a constraint-based algorithm that leverages the natural causal precedence of genotypes to infer ancestral relationships in transcriptomic data. Unlike traditional causal discovery methods that start with a fully connected graph, GENESIS initialises an empty ancestrality matrix and iteratively populates it with direct, indirect or non-causal relationships using a series of provably sound marginal and conditional independence tests. By integrating genotypes as fixed causal anchors, GENESIS provides a principled ``head start'' to classical causal discovery algorithms, restricting the search space to biologically plausible edges. We test GENESIS on synthetic and real-world genomic datasets. This framework offers a powerful avenue for uncovering causal pathways in complex traits, with promising applications to functional genomics, drug discovery, and precision medicine.

Kirtan Padh, Jakob Zeitler, David Watson et al.

Causal effect estimation is important for many tasks in the natural and social sciences. We design algorithms for the continuous partial identification problem: bounding the effects of multivariate, continuous treatments when unmeasured confounding makes identification impossible. Specifically, we cast causal effects as objective functions within a constrained optimization problem, and minimize/maximize these functions to obtain bounds. We combine flexible learning algorithms with Monte Carlo methods to implement a family of solutions under the name of stochastic causal programming. In particular, we show how the generic framework can be efficiently formulated in settings where auxiliary variables are clustered into pre-treatment and post-treatment sets, where no fine-grained causal graph can be easily specified. In these settings, we can avoid the need for fully specifying the distribution family of hidden common causes. Monte Carlo computation is also much simplified, leading to algorithms which are more computationally stable against alternatives.

David Watson, Limor Gultchin, Ankur Taly et al.

Necessity and sufficiency are the building blocks of all successful explanations. Yet despite their importance, these notions have been conceptually underdeveloped and inconsistently applied in explainable artificial intelligence (XAI), a fast-growing research area that is so far lacking in firm theoretical foundations. Building on work in logic, probability, and causality, we establish the central role of necessity and sufficiency in XAI, unifying seemingly disparate methods in a single formal framework. We provide a sound and complete algorithm for computing explanatory factors with respect to a given context, and demonstrate its flexibility and competitive performance against state of the art alternatives on various tasks.

David Kinney, David Watson

Discovering high-level causal relations from low-level data is an important and challenging problem that comes up frequently in the natural and social sciences. In a series of papers, Chalupka et al. (2015, 2016a, 2016b, 2017) develop a procedure for causal feature learning (CFL) in an effort to automate this task. We argue that CFL does not recommend coarsening in cases where pragmatic considerations rule in favor of it, and recommends coarsening in cases where pragmatic considerations rule against it. We propose a new technique, pragmatic causal feature learning (PCFL), which extends the original CFL algorithm in useful and intuitive ways. We show that PCFL has the same attractive measure-theoretic properties as the original CFL algorithm. We compare the performance of both methods through theoretical analysis and experiments.