Daniel E. Ho
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Daniel E. Ho has a total of 173 co-authors. Here are the top co-authors by number of shared publications:
- Derek Ouyang
Regulation, Evaluation, and Governance Lab, Stanford University, Stanford, CA, USA
12 shared publications - Jacob Goldin
Law School, University of Chicago, Chicago, Illinois, USA, and American Bar Foundation, Chicago, Illinois, USA
5 shared publications - Kit T. Rodolfa
Stanford University, Stanford, USA
5 shared publications Sarah L. Rudman
County of Santa Clara Public Health Department, San Jose, CA 95126, USA
5 shared publications- Christopher D. Manning
Department of Computer Science, Stanford University
4 shared publications Alexis D’aGostino
County of Santa Clara Public Health Department, San Jose, CA 95126, USA
3 shared publications
Latest Publications
- 21 September 2026
- Published by Wolters Kluwer Health in Epidemiology
Association Between Proximity to Animal Agriculture During Pregnancy and Spontaneous Preterm Birth Among Nulliparous Women in a California Cohort
Background: Environmental exposures during pregnancy may influence risk of spontaneous preterm birth, a serious adverse outcome. Concentrated Animal Feeding Operations (CAFOs) near maternal residences may emit pollutants associated with preterm birth. Methods: We analyzed n = 811,953 births to nulliparous women in California (2007–2011), including 7487 spontaneous very preterm (20–31 weeks) and 37,026 spontaneous moderate preterm (32–36 weeks) births. Locations of n = 2121 CAFOs were obtained from a near-census of facilities. Using time-to-event and Bayesian methods, we estimated hazard and odds of preterm birth as a function of distance, density, and wind-based CAFO exposure, adjusting for maternal, temporal, socioeconomic, and environmental covariates. Bayesian models allow inference on both the distance range of association and the association strength. Results: Greater exposure to CAFOs is associated with small increases in hazard of all preterm birth (adjusted hazard ratios, per-unit exposure change − inverse distance: 1.003 [95% confidence interval [CI] = 1.000, 1.005]; downwind-weighted: 1.020 [95% CI = 1.005, 1.035]) with stronger associations for very preterm birth (adjusted hazard ratios − inverse distance: 1.007 [95% CI = 1.002, 1.012]; downwind-weighted: 1.043 [95% CI = 1.008, 1.080]). Among exposed births, less downwind exposure is associated with lower preterm birth hazard compared with being predominantly downwind. Bayesian models suggest births with at least one CAFO within 25.23 km (95% highest density interval [HDI]: 18.21, 32.74 km) have 1.16 times the odds of being very preterm (95% HDI: 1.10, 1.22), with similar distance ranges but smaller odds for moderate preterm birth (adjusted odds ratio: 1.08 [95% HDI: 1.05, 1.10]). Conclusions: Living near a CAFO may increase risk of spontaneous preterm birth in California, with stronger and more consistent associations for very preterm births, particularly when combined with greater downwind exposure.
- 21 September 2026
- Published by Wolters Kluwer Health in Epidemiology
Association Between Proximity to Animal Agriculture During Pregnancy and Spontaneous Preterm Birth Among Nulliparous Women in a California Cohort
Background: Environmental exposures during pregnancy may influence risk of spontaneous preterm birth, a serious adverse outcome. Concentrated Animal Feeding Operations (CAFOs) near maternal residences may emit pollutants associated with preterm birth. Methods: We analyzed n = 811,953 births to nulliparous women in California (2007–2011), including 7487 spontaneous very preterm (20–31 weeks) and 37,026 spontaneous moderate preterm (32–36 weeks) births. Locations of n = 2121 CAFOs were obtained from a near-census of facilities. Using time-to-event and Bayesian methods, we estimated hazard and odds of preterm birth as a function of distance, density, and wind-based CAFO exposure, adjusting for maternal, temporal, socioeconomic, and environmental covariates. Bayesian models allow inference on both the distance range of association and the association strength. Results: Greater exposure to CAFOs is associated with small increases in hazard of all preterm birth (adjusted hazard ratios, per-unit exposure change − inverse distance: 1.003 [95% confidence interval [CI] = 1.000, 1.005]; downwind-weighted: 1.020 [95% CI = 1.005, 1.035]) with stronger associations for very preterm birth (adjusted hazard ratios − inverse distance: 1.007 [95% CI = 1.002, 1.012]; downwind-weighted: 1.043 [95% CI = 1.008, 1.080]). Among exposed births, less downwind exposure is associated with lower preterm birth hazard compared with being predominantly downwind. Bayesian models suggest births with at least one CAFO within 25.23 km (95% highest density interval [HDI]: 18.21, 32.74 km) have 1.16 times the odds of being very preterm (95% HDI: 1.10, 1.22), with similar distance ranges but smaller odds for moderate preterm birth (adjusted odds ratio: 1.08 [95% HDI: 1.05, 1.10]). Conclusions: Living near a CAFO may increase risk of spontaneous preterm birth in California, with stronger and more consistent associations for very preterm births, particularly when combined with greater downwind exposure.
- Open Access
- 20 July 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (30)
At the boundary: Law and AI
- 20 July 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (30)
Open AccessAt the boundary: Law and AI
- Open Access
- 20 July 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (30)
There is no free benchmark: An institutional view of legal AI benchmarking
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.
- 20 July 2026
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 123 (30)
Open AccessThere is no free benchmark: An institutional view of legal AI benchmarking
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.
- Open Access
- 10 July 2026
- Published by Cambridge University Press (CUP) in Political Analysis
- p. 1-20
Monotone Ecological Inference
We study the identification of individual-level associations when only aggregate data are available. We characterize the biases of, and relationships among, canonical ecological inference (EI) estimators. We use these results to develop a partial identification approach: monotone EI. The approach exploits information about one or both of the following conditional associations: (1) outcome differences between groups within the same neighborhood and (2) outcome differences within the same group between neighborhoods with different group compositions. We show how assumptions about the sign of these conditional associations, whether individually or in relation to one another, can yield informative sharp bounds. We illustrate our results using county-level data to study differences in COVID-19 vaccination rates among Republicans and Democrats in the United States.
- 10 July 2026
- Published by Cambridge University Press (CUP) in Political Analysis
- p. 1-20
Open AccessMonotone Ecological Inference
We study the identification of individual-level associations when only aggregate data are available. We characterize the biases of, and relationships among, canonical ecological inference (EI) estimators. We use these results to develop a partial identification approach: monotone EI. The approach exploits information about one or both of the following conditional associations: (1) outcome differences between groups within the same neighborhood and (2) outcome differences within the same group between neighborhoods with different group compositions. We show how assumptions about the sign of these conditional associations, whether individually or in relation to one another, can yield informative sharp bounds. We illustrate our results using county-level data to study differences in COVID-19 vaccination rates among Republicans and Democrats in the United States.
- Open Access
- 2 July 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
How Defining ‘Disadvantaged’ Shaped Bipartisan Infrastructure Law Drinking Water Funding
The 2021 Bipartisan Infrastructure Law (BIL) directed $30.7 billion specifically towards safe drinking water programs, representing the single-largest investment into water infrastructure in United States history. The law mandates that 49% of funds be provided as grants and forgivable loans to "disadvantaged communities." Notably, the definition of disadvantaged communities has been left to the discretion of states, allowing variability in how different states identify and prioritize these communities. This study examines how states redefined disadvantaged communities after the BIL, maps their spatial distribution, and analyzes disparities in drinking water funding from 2019 to 2023. Our results suggest that, despite widespread updates to disadvantaged community definitions, racial and ethnic minorities continue to encounter barriers to accessing funding. If policies are aimed to address historical disparities in drinking water investment, definitional delegation appears to fall short, and ongoing engagement with underserved communities may be needed for more equitable infrastructure investment.
- 2 July 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Open AccessHow Defining ‘Disadvantaged’ Shaped Bipartisan Infrastructure Law Drinking Water Funding
The 2021 Bipartisan Infrastructure Law (BIL) directed $30.7 billion specifically towards safe drinking water programs, representing the single-largest investment into water infrastructure in United States history. The law mandates that 49% of funds be provided as grants and forgivable loans to "disadvantaged communities." Notably, the definition of disadvantaged communities has been left to the discretion of states, allowing variability in how different states identify and prioritize these communities. This study examines how states redefined disadvantaged communities after the BIL, maps their spatial distribution, and analyzes disparities in drinking water funding from 2019 to 2023. Our results suggest that, despite widespread updates to disadvantaged community definitions, racial and ethnic minorities continue to encounter barriers to accessing funding. If policies are aimed to address historical disparities in drinking water investment, definitional delegation appears to fall short, and ongoing engagement with underserved communities may be needed for more equitable infrastructure investment.
- Open Access
- 25 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 7019-7053
Privacy Without Remedy: An Assessment of Data Broker Compliance with California Privacy Law
- 25 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 7019-7053
Open AccessPrivacy Without Remedy: An Assessment of Data Broker Compliance with California Privacy Law
- Open Access
- 25 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 3322-3348
Tradeoffs are Domain Dependent: Improving Accuracy and Fairness in Property Tax Assessments
- 25 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 3322-3348
Open AccessTradeoffs are Domain Dependent: Improving Accuracy and Fairness in Property Tax Assessments
- Open Access
- 14 May 2026
- Published by Elsevier in The Lancet Regional Health - Americas
- Vol. 60, 101495
Partisan differences in childhood measles vaccination and general refusals: a retrospective cohort study of electronic health records in the United States, 1988–2024
- 14 May 2026
- Published by Elsevier in The Lancet Regional Health - Americas
- Vol. 60, 101495
Open AccessPartisan differences in childhood measles vaccination and general refusals: a retrospective cohort study of electronic health records in the United States, 1988–2024
- Open Access
- 30 April 2026
- Published by MIT Press in Harvard Data Science Review
- Vol. 8 (2)
Explainability in the Wild, or Wild Explanations? Evidence From Predicting Tax Evasion
As artificial intelligence algorithms become more prevalent in high-stakes risk assessment, policymakers have increasingly relied on explainability tools for interpretability. Despite growing mandates that AI-based decisions include explanations, there remains little empirical evidence demonstrating the effectiveness of these techniques in real-world applications. This gap often stems from the absence of a clear ground truth for evaluating explanations. In this work, we present an empirical evaluation of explanation techniques in collaboration with the United States Internal Revenue Service (IRS). Using real, line-by-line IRS audits from randomly-selected taxpayers, we decompose one component of aggregate tax under-reporting into its constituent line-item misreporting and apply explainability techniques to recover these risks; the aggregate risk is a function of the constituent risk. Our study makes three contributions. First, we empirically evaluate how well local explanation models recover true constituent risks. Second, we compare local explanation models to estimating constituent risks directly. Finally, we situate these findings in a practical setting where explanations are critical not only for transparency but also as guidance for the users of the model's predictions. Our analysis reveals that the quality of local explanations is tied to the quality of the underlying model. Yet even with a theoretically perfect underlying model, local explanations still fail to accurately capture the true risk. While directly estimating constituent risks may yield more accurate results, simplistic rule-based heuristics often overlook the complexity of risk. These findings highlight the need for thoughtful application of explanation techniques in high-risk domains, where errors can have significant consequences.
- 30 April 2026
- Published by MIT Press in Harvard Data Science Review
- Vol. 8 (2)
Open AccessExplainability in the Wild, or Wild Explanations? Evidence From Predicting Tax Evasion
As artificial intelligence algorithms become more prevalent in high-stakes risk assessment, policymakers have increasingly relied on explainability tools for interpretability. Despite growing mandates that AI-based decisions include explanations, there remains little empirical evidence demonstrating the effectiveness of these techniques in real-world applications. This gap often stems from the absence of a clear ground truth for evaluating explanations. In this work, we present an empirical evaluation of explanation techniques in collaboration with the United States Internal Revenue Service (IRS). Using real, line-by-line IRS audits from randomly-selected taxpayers, we decompose one component of aggregate tax under-reporting into its constituent line-item misreporting and apply explainability techniques to recover these risks; the aggregate risk is a function of the constituent risk. Our study makes three contributions. First, we empirically evaluate how well local explanation models recover true constituent risks. Second, we compare local explanation models to estimating constituent risks directly. Finally, we situate these findings in a practical setting where explanations are critical not only for transparency but also as guidance for the users of the model's predictions. Our analysis reveals that the quality of local explanations is tied to the quality of the underlying model. Yet even with a theoretically perfect underlying model, local explanations still fail to accurately capture the true risk. While directly estimating constituent risks may yield more accurate results, simplistic rule-based heuristics often overlook the complexity of risk. These findings highlight the need for thoughtful application of explanation techniques in high-risk domains, where errors can have significant consequences.
- Top Cited PapersOpen Access
- 3 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 101-125
Benchmarking Legal RAG: The Promise and Limits of AI Statutory Surveys
- 3 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 101-125
Top Cited PapersOpen AccessBenchmarking Legal RAG: The Promise and Limits of AI Statutory Surveys