There is no free benchmark: An institutional view of legal AI benchmarking.

July 28, 2026| Source: Proceedings of the National Academy of Sciences of the United States of America

Despite substantial excitement around the use of AI in law, little information exists on the performance and associated risks of the domain's widely marketed tools. Recent work, for instance, has demonstrated the significant potential for "hallucinations"-wherein models make up facts, law, and precedent-leading Chief Justice Roberts to spotlight this risk in his annual report on the judiciary. We argue that there is a need for public AI benchmarking in law. First, relative to other AI application domains, the legal AI ecosystem lacks legibility-there is little information about the design and performance of many commercial legal AI systems. Legal AI has not benefited from the types of benchmarking that have catalyzed, measured, and informed AI innovation and responsible use in other domains. Second, we articulate the challenges of the institutional design of benchmarking. We illustrate how benchmarks can be captured, watered down, and abused. Careful institutional design around the why, who, what, and how of benchmarking will be critical to navigate difficult tradeoffs of transparency, objectivity, expertise, and resources. Third, addressing legal AI's illegibility requires matching institutional models to available resources and constraints. Rather than advocating for a single "best" approach to benchmarking, we show how benchmarking strategies depend on available resources.

Read the original source →

Nature's antivenom: Combinations of conserved rattlesnake serum metalloproteinase inhibitors block the lethal action of viper venoms.

Snakebite maims or kills several hundred thousand people each year. For more than a century, treatment has relied on antivenoms derived from animals immunized with whole venoms, but their efficacy, safety, and availability are highly variable, and it is often not well understood which specific venom components must be inhibited to prevent mortality and major morbidities. New therapeutic approaches are needed. Here, we take an evolutionary approach to antivenom design inspired by the longstanding

Physical/TechAugust 4, 2026

Abrupt transition to irreversible damage in the overdrafted Sacramento Valley aquifer system.

Groundwater extraction decreases water pressure in aquifer systems, causing reversible or irreversible deformation of the water-bearing layers that manifests as recoverable or permanent displacements of the land surface, respectively. Detecting and forecasting when and where an aquifer system transitions from a reversible, poroelastic regime, to an irreversible, inelastic regime remains a crucial challenge given the complex, heterogeneous nature of aquifer systems. Here we leverage high-resoluti

Physical/TechJuly 28, 2026

Synergistic action of different molecular mechanisms causes striking levels of insecticide resistance in the malaria vector Anopheles gambiae .

Intensifying insecticide resistance in the malaria vector Anopheles gambiae poses a serious threat to the progress achieved the last decades in reducing malaria deaths in Africa. The genetic basis of insecticide resistance is often complex, involving multiple genes and mutations. However, we still lack a clear understanding of how each mechanism contributes to overall resistance and how highly resistant phenotypes arise. In this study, we generated a suite of transgenic An. gambiae strains carry

Physical/TechJuly 28, 2026

A phosphorylation cascade involving ZmSnRK2.10-ZmRIPK2-ZmWRKY38 attenuates drought response by derepressing ZmSUS2 in maize.

Drought stress severely limits crop productivity, with transcription factors (TFs) playing pivotal roles in plant adaptation. Here, we identify the maize TF ZmWRKY38 as a positive regulator of drought tolerance. CRISPR/Cas9-mediated knockout of ZmWRKY38 increased plant sensitivity to drought compared with the wild type. We demonstrate that ZmWRKY38 binds to the promoter of the sucrose synthase gene ZmSUS2 and represses its transcription. Conversely, ZmSUS2 functions as a negative regulator of dr

Physical/TechJuly 28, 2026

The quantum ensemble variational optimization algorithm: Applications to molecular inverse design.

Designing molecules with optimized properties remains a fundamental challenge due to the intricate relationship between molecular structure and properties. Traditional computational approaches that address the combinatorial number of possible molecular designs become unfeasible as the molecular size increases, suffering from the so-called "curse of dimensionality" problem. Recent advances in quantum computing hardware present new opportunities to address this problem. Here, we introduce the quan