This report presents findings from the fourth National Inmate Survey in state and federal prisons as part of the federal Bureau of Justice Statistics’ National Prison Rape Statistics Program, which collects allegations of sexual victimization directly from victims. The report describes characteristics of victims, perpetrators, and incidents related to inmate-reported sexual victimization in prisons. Findings include that among female inmates, a larger percentage of Black inmates (5.6%) reported inmate-on-inmate victimization than White inmates (2.9%), but among male inmates, a larger percentage of White inmates (2.9%) reported inmate-on-inmate victimization than Black inmates (1.1%). A smaller percentage of female inmates (4.1%) reported inmate-on-inmate victimization in 2023–24 than 2011–12 (6.9%), but a larger percentage of male inmates (2.2%) reported inmate-on-inmate victimization in 2023–24 than 2011–12 (1.7%). A greater percentage of inmates who were serving in the U.S. Armed Forces at the time of the survey reported sexual victimization of any type (15.2%), inmate-on-inmate victimization (10.3%), or staff-on-inmate victimization (9.4%) than inmates who had never served. Thirty-one percent of inmates who reported prior sexual victimization in another facility also reported sexual victimization in their current facility.

In 2026, 425 criminal justice laws passed across 45 states—every state with an active legislative session except Arkansas—and Washington, D.C. Democratic-led states passed 58% of all reform laws, Republican-led states passed 28%, and divided states passed 14%. Reentry laws were enacted more than any other type, including Maryland’s Fair Chance Housing Act, Louisiana’s transitional reentry program for incarcerated women, and Nebraska’s Youth Reentry and Transitional Support Act. Juvenile justice reform also took major steps forward in 2026 with 33 laws passed. Nine states restricted state and local law enforcement cooperation with U.S. Immigration and Customs Enforcement, while six states empowered residents to sue federal agents for constitutional violations. Surveillance and policing were thrust into the spotlight this year after a rapid expansion in surveillance tech and swift public backlash. Five states, including divided Kentucky, enacted new guardrails on automated license plate readers. Florida passed six criminal justice focused laws in 2026 and also took action to restrict police use of Flock surveillance systems.

Many public school facilities are outdated or poorly maintained. Facility issues, including plumbing leaks and extreme indoor temperatures, can disrupt instruction or even result in cancelled classes. According to a nationally representative survey of public school districts (n=765), chronic facility issues disrupted teaching and learning in schools nationwide in school year 2024-25, with nearly one in five districts (19%) canceling school due to a facilities issue. These cancellations affected approximately 2 million students, who lost an estimated 3.5 million days of learning as a result. Top reasons for canceling school included plumbing emergencies and extreme classroom temperatures. For example, one district reported canceling school after classroom temperatures reached 110 degrees. Nearly half of districts (47%) had chronic facility issues that often disrupted learning. Findings also include that districts frequently face funding constraints and have used strategies like investing in preventive maintenance to manage costs and better meet facility needs, though 55% of responding districts considered their most recent facilities budget insufficient to meet their needs. According to the authors, districts with limited capacity to conduct preventive maintenance had more chronic facility problems and more disruptions to instruction than other districts.

Generative artificial intelligence (AI) has been promoted as the technology that could transform education by providing every student a personal tutor. This study provides the first large-scale experimental evidence of the effect of providing students with an AI tutor on math scores using a two-year randomized control trial in 18 Tennessee middle schools. Students in the treatment condition were randomly assigned to use Khan Academy with its AI tutor, Khanmigo, which was configured to coach rather than give answers, while students in the control condition received usual remedial math work. Findings include that Khan Academy with Khanmigo raised math achievement by 1.3 national percentile ranks per term, or about 0.06 to 0.08 standard deviations over a school year. These gains resemble those of studies of Khan Academy without AI assistance. The authors note that nearly all students in the treatment condition tried Khanmigo, often attempting to use it to get answers rather than assistance or coaching. However, when Khanmigo instead asked questions or offered hints, students largely stopped using it. The authors conclude that seeking help with one’s own confusion remained a choice, and most students in the experiment declined it most of the time.

Graduate schools of education (GSEs) are among the most influential institutions in American K–12 schooling, yet there is little evidence that attendance at such an institution does much to improve teacher quality. To understand what is taught at GSEs, the author of this report analyzed one master’s or certification program that aligned most closely with an elementary teaching degree at eight public GSEs across five states, including one program at the University of Florida (UF) and one at Florida International University (FIU). Readings from 75 course syllabi were sorted into four categories of educational philosophy: traditional, progressive, critical, and therapeutic. Per the author, several programs displayed pronounced ideological asymmetry, with critical and progressive orientations substantially outweighing traditional or practical materials in the accessible course records. Only in the FIU program were more readings classified as traditional—focused on topics like phonics or classroom management, or classic books—than in any other category. According to the author, findings raise questions about whether GSEs are preparing teachers for the realities of the classroom or are more engaged in a project of shaping prospective teachers’ worldviews.

The impact of artificial intelligence (AI) on the employment prospects of recent college graduates is hotly debated with no consensus on the magnitude or timing of potential impacts. Using Current Population Survey microdata, the authors provide the first estimates of the effects of AI on the unemployment of recent college graduates in June, July, and August 2026. Key findings include that unemployment rates did not spike in summer 2026 relative to summer months in previous years and did not rise in a significant way relative to older college graduates or young workers without a college degree. Under an expanded definition of “unemployed” that includes those who report wanting a job, the authors still found no evidence of a statistically significant increase in unemployment among recent college graduates in summer 2026. The authors conclude that there is currently no evidence of any significant, widespread displacement or reduction in hiring of recent college graduates in absolute or relative levels.

In 2024, 84% of federal revenue came from either individual taxes or payroll taxes, both of which are collected from labor income. This report examines the fiscal implications for the U.S. federal government in scenarios in which artificial intelligence (AI) and automation technologies replace human labor. Using a framework that considers whether displaced workers find new jobs and how AI is priced, the authors analyze a hypothetical case of full displacement of human labor and a transitional scenario in which AI displaces 10% of the workforce. Findings include that because of the concentration of income at the top of the distribution in high-paid, white-collar jobs, the replacement of these roles by AI could disrupt the majority of federal income tax revenue. In addition, the loss of jobs and income would reduce economic activity in general because most households rely on labor income to support consumption. This reduction in economic activity could adversely affect businesses, which would further reduce federal and state revenues. The authors make several recommendations, including changing the tax code to reduce dependence on labor income, planning income support if labor income falls sharply, and regulating AI systems if windfall profits emerge.

This report examines 21 projects funded by Washington state’s Climate Commitment Act (CCA), a market-based cap-and-invest program created in 2021 to reduce greenhouse gas emissions. Per the author, CCA grant dollars were to be used to reduce air pollution in overburdened communities. Key findings include that the majority of projects funded by the CCA produced little or no reduction in air pollution. Moreover, 58% of CCA project funds went to projects’ staffing, 10% to overhead, and only 13% to equipment that could directly impact air quality. Other findings include that no CCA-funded project will collect environmental monitoring data, making it impossible to verify intervention effectiveness; projects were funded without requirements for measurable results, cost-effectiveness, or consequences for failure; and many of the grants were used for “policy advocacy” and organizing efforts rather than direct pollution reduction. The author notes that only one of the twenty-one funded projects stood out as cost effective, a Spokane road-paving project that both quantified pollution reduction and delivered it at a reasonable cost per ton ($1,400).

On a given day in 2025, an estimated 1 million residents lived in approximately 31,400 residential care communities. Many of these communities served residents with multiple chronic conditions, indicating potential complex needs among their resident populations. This report uses 2025 data to examine the characteristics of residential care communities that reported four chronic conditions among their current residents: Alzheimer disease and other dementias, heart disease, diabetes, and depression. Key findings from the report include that in 2025, almost one-half of residential care communities reported having residents diagnosed with all four selected chronic conditions (48.0%)—Alzheimer disease or other dementias, diabetes, depression, and heart disease. Residential care communities with 11 or more beds and those located in nonmetropolitan areas, the Northeast, and the Midwest were more likely to report having all four chronic conditions represented among their current residents. A higher percentage of nonprofit (54.3%) and chain-affiliated (54.3%) residential care communities, as well as communities using electronic health records (56.7%), reported having all four chronic conditions represented. Residential care communities that reported adverse events among their current residents (emergency department visits, hospitalizations, and falls) showed a higher prevalence of all four conditions.

This study analyzes material hardship among young adults ages 18 to 24 using data from the Well-Being and Basic Needs Survey. Other findings include that food insecurity was the most commonly reported hardship among young adults (30%) in 2025, and rates of material hardship among young adults were higher than those among adults ages 25 to 64 across the following: household food insecurity, problems paying rent or mortgage, and utility shutoffs. Although young adults who are eligible for the federal Supplemental Nutrition Assistance Program (SNAP) and housing assistance participated in both at lower rates than adults ages 25–64, their households receive charitable food at higher rates, with almost one in four young adults reporting receiving charitable food in 2025. The authors suggest that many young adults enter a pivotal life stage with higher levels of hardship and lower access to the safety net than other adults, even when they are eligible for assistance.

Approximately 80% of U.S. adults take at least one supplement for their health according to a nationally representative (n=3,554) survey by the Pew Research Center. Vitamins and minerals are by far the most common supplement, with two-thirds of survey respondents (67%) saying they currently take a vitamin or mineral supplement for their health. Findings include that older adults are more likely than younger adults to take supplements, with 85% of respondents ages 65 and older taking at least one compared to 67% of adults under 30. Age differences account for some differences in supplement use. For example, most adults 65 and older (81%) take vitamins or minerals, compared with 48% of adults under 30. However, younger adults are more likely to take protein supplements than those ages 65 and older, with a little over one-third of adults ages 18 to 29 (37%) and those ages 30 to 49 (35%) taking protein supplements compared to 23% of adults 65 and older. Most adults who take supplements perceive them as important for health. For example, among those who take vitamins or minerals, 62% say they are extremely or very important to their health while another 33% say they are somewhat important.