It is argued that meaningful assessment must go beyond metric‐centric model evaluation to encompass biological realism, data diversity, confounding control, and interpretability, and a unifying framework, VALID, comprising the following five guiding principles for rigorous AI in biology is proposed.
Abstract
Abstract Artificial intelligence (AI) is rapidly transforming computational biology. However, closer scrutiny reveals that many reported results do not hold under rigorous evaluation. This paper identifies prevalent, recurring issues in the field, including narrow datasets, hidden confounders that make supposedly independent test data effectively indistinguishable from the training data (the doppelgänger effect), misleading performance metrics, and over‐reliance on overly easy test cases. We argue that meaningful assessment must go beyond metric‐centric model evaluation to encompass biological realism, data diversity, confounding control, and interpretability. Crucially, AI studies in computational biology should deliver not only predictive performance but also genuine biological insight and transferable methodological understanding. To address these systemic limitations, we propose a unifying framework, VALID, comprising the following five guiding principles for rigorous AI in biology: Veracious biological grounding and novelty, artifact control and methodological rigor, legitimate generalization, inspectable preprocessing, and deployment‐aware performance. Together, these principles serve as a diagnostic lens for structuring our discussion throughout this paper. We further illustrate common pitfalls through representative examples and provide actionable guidance for authors, reviewers, and editors aimed at improving the rigor and scientific value of AI‐driven computational biology.
Abstract Motivation Effective computational tool development depends on our ability to measure progress with rigorous benchmarking, allowing developers to showcase advances and users to select the best of the available options. Computational biology has developed a robust infrastructure of benchmarking standards, tools...
Russell Schwartz, Xue-Gong Zhang, Jian Ma et al.· Bioinformatics Advances· 1 citation
Reliable automated research requires agents to vet data, verify analyses, and generate hypotheses grounded in trustworthy evidence, potentially reducing routine scientific workload while allowing scientists to focus on interpretation and discovery. Existing benchmarks often only assess analytical task completion or hyp...
Nan Huang, Mario Tapia-Pacheco, Kun Zhou et al.· arXiv.org· 0 citations
Artificial intelligence is increasingly used across chemical research, but its practical value depends on where it intervenes in the workflow and how reliably it generalizes beyond curated benchmarks. This review adopts a pipeline-based perspective: we first decompose chemical research into stages that generate, transf...
M. Medvedev, Nikita S. Serov, V. Chaliy et al.· Chemical Society Reviews· 0 citations
Neural networks are likely to be among the most consequential artefacts in human history. As these systems become more capable and autonomous, their potential to cause societal harm also grows, making AI safety one of the defining challenges of our time. Yet we continue to build increasingly powerful systems without fu...
Fazl Barez, Tung-Yu Wu, Hunar Batra et al.· 0 citations
Deep learning and ensemble models now drive decisions in healthcare, finance, and judicial systems, yet their internal
reasoning remains largely opaque to the humans who rely on them. This opacity — the "black-box problem" — limits trust,
complicates regulatory compliance, and makes it difficult to detect bias or error...
Varshit Khaliya, Sudhanshu Ranjan, Vishal Shrivastava et al.· International Journal for Re...· 0 citations
AI-driven discovery does not create the historical fact that results can precede explanation. It changes how that interval is produced and governed. Systems such as AlphaFold, GNoME, and FunSearch can generate high-value outputs at a rate and scale that require communities to allocate attention, validation resources, a...