Washington’s research shops spent the first week of August publishing on the two subjects that increasingly crowd out everything else — artificial intelligence and biological risk — while continuing a steady output on immigration, health care pricing, and the fiscal treatment of capital investment. The Investigative Journal reviewed publications released between July 27 and August 6, 2026, across the ideological spectrum. Below are eight that merit attention, with each institution’s orientation and funding base noted so readers can weigh the source alongside the finding.
1. RAND: A preemptive case against “mirror life”
The week’s most consequential release came from the RAND Corporation, which published A U.S. Strategy to Prevent the Creation of Mirror Life on July 31. The 69-page report addresses synthetic organisms constructed from mirror-image biomolecules — reversed chirality relative to all known life. RAND’s authors, a ten-person team led by Nidhi Kalra, argue that such organisms could evade immune detection, resist normal biological degradation, and escape natural predation, and that specialists in the field assess creation to be technically feasible within decades.
The strategic argument is the notable part. RAND contends that the governance model the United States typically applies to emerging technology — watch, assess, then regulate — fails structurally in this case, because there are few observable early-warning signals and the first successful experiment may be the point at which containment is no longer possible. The report accordingly recommends that Washington publicly commit never to develop mirror life even if it suspects rivals are pursuing it, and engage China and other scientific powers directly rather than assembling an allied-bloc regime that could itself provoke a race. RAND released a companion risk-communication paper, available here, the same day.
RAND is a nonpartisan federally chartered research organization and reports that the majority of its research revenue comes from U.S. government contracts and grants, principally defense and health agencies, supplemented by philanthropic support. Readers should note that a recommendation of unilateral restraint is an unusual posture for an institution whose primary client is the Department of Defense.
2. Council on Foreign Relations: experts say AI governance is failing
CFR published How Experts Think AI Will Change the Global Balance of Power by 2035 on July 30, based on a survey fielded April 30 to June 16, 2026, drawing 362 responses from CFR members and the Council of Councils network, with 237 completing all sections.
The headline numbers are sobering. More than 80 percent of respondents rated expected AI governance coherence below 50 on a 0-to-100 scale, with a mean of 32 and median of 29. Only 5 percent said domestic institutions are keeping pace with the technology. Expectations about capability concentration were sharply bimodal — 46 percent expect frontier AI to remain held by two or three actors, 54 percent expect diffusion, with only 10 percent clustering near the midpoint. Nearly 70 percent said frontier AI laboratories will be the most powerful nonstate actors by 2035, and more than 70 percent identified a serious AI accident as the likeliest trigger for coherent governance. A generational split is visible in the data: 61 percent of respondents aged 35 to 49 expect concentration, against 32 percent of those over 65.
CFR is nonpartisan and funded principally through membership dues, a corporate program, foundation grants, and its endowment. The survey questions and underlying data are downloadable, which permits independent replication — a transparency practice more think tanks should adopt.
3. AEI and CSIS: putting China’s carriers at risk
The American Enterprise Institute, jointly with the Center for Strategic and International Studies, published Flipping the Script: How to Hold China’s New Carriers at Risk on August 4. Authors Mackenzie Eaglen, Benjamin Jensen, Henry H. Carroll, and Harold “Bud” Bishop invert the usual framing: rather than treating the People’s Liberation Army Navy’s carrier expansion as a threat to be matched, they treat it as a capital-intensive vulnerability to be exploited through cost imposition.
The report proposes five reinforcing lines of effort — forward subsurface ambush forces, distributed land-based maritime strike, long-range air and space-enabled fires, cyber and electromagnetic attack, and unmanned swarm and port-denial operations. Its force-structure recommendation is a fleet weighted toward roughly 300 crewed and 300 uncrewed vessels, alongside expanded torpedo and undersea drone production, strengthened Marine and Army littoral forces at maritime chokepoints, and missile co-production with Australia and Japan.
AEI is a center-right institution that states it accepts no government funding, drawing instead on individual donors, corporations, and foundations; the Bradley Foundation and Searle Freedom Trust have been among its reported foundation supporters. CSIS is bipartisan and discloses corporate, foundation, and government funding, including from foreign governments — a disclosure worth weighing when a report recommends co-production with named allied states.
4. Cato Institute: citizenship and youth crime in the German natural experiment
On August 5, the Cato Institute published Birthright Citizenship and Youth Crime as Research Brief No. 495, summarizing work by Leander Andres, Stefan Bauernschuster, Gordon B. Dahl, Helmut Rainer, and Simone Schüller circulated as NBER Working Paper 35070.
The study exploits Germany’s January 1, 2000 citizenship reform, which raised the share of second- and third-generation immigrant children born with citizenship from 22 percent to 89 percent. Comparing birth cohorts immediately before and after the cutoff, the authors report a roughly 70 percent reduction in youth crime among children who gained citizenship through the reform, concentrated in property and drug offenses and driven largely by boys. The data cover all police-reported crime in Baden-Württemberg, Berlin, and Hesse — some 20.5 million residents, about a quarter of Germany’s population — for offenses recorded between July 2012 and February 2020.
The findings speak to an active American legal and policy debate, though the German institutional context differs materially from the U.S. one and the authors’ identification strategy rests on a single national reform. Cato is libertarian, was founded in 1977 with backing from Charles Koch, and states that it accepts no government funding, drawing most revenue from individual donors.
5. Manhattan Institute: voters back work requirements thirty years after welfare reform
Marking three decades since the 1996 welfare reform, the Manhattan Institute released American Attitudes on Welfare, Waste, Crime, and the Political Environment on July 27. Jesse Arm and Karl Abramson surveyed 1,121 likely 2026 midterm voters between July 1 and 6, combining a 701-respondent online panel with 420 SMS-to-web interviews, weighted to age, race, gender, education, and 2024 vote, with a reported margin of error of ±3.1 points.
On social policy the results show durable consensus: 74 percent support work requirements, 80 percent prefer in-kind benefits to cash transfers, and voters oppose a universal basic income 50 to 32 — though respondents under 35 support it 49 to 30. Fifty-seven percent back the 2025 expansion of SNAP work requirements, and 82 percent oppose abolishing police departments. On the political environment, the poll finds presidential job approval at 44 percent against 54 percent disapproval, Democrats leading the generic congressional ballot 47 to 43, and a 2028 Republican primary field in which JD Vance leads Marco Rubio 37 to 20 — a lead the poll indicates would reverse to 43-36 for Rubio under a hypothetical presidential endorsement.
The Manhattan Institute is right-of-center and is supported by individual donors and foundations, with the Bradley Foundation and Searle Freedom Trust among its reported backers. As with any single-sponsor poll, the horse-race numbers should be read against other public surveys rather than in isolation.
6. Roosevelt Institute: a blueprint for 2029
The Roosevelt Institute published Building Up in 2029: How to Make Green Statecraft Durable on July 30 — a 17-essay compendium with contributions from K. Sabeel Rahman, Heather Boushey, Noah Kaufman, Madison Condon, Shelley Welton, Saule Omarova, Lenore Palladino, and Leah C. Stokes, among others.
The volume is organized around four questions — how to govern, plan, finance, and produce climate and industrial policy — and its proposals are institutional rather than incremental: a National Climate Service, a sovereign wealth vehicle modeled on the Reconstruction Finance Corporation, public equity stakes in strategic firms, publicly owned production capacity, and Senate procedural reform. The introduction frames the project explicitly as a structural answer to Project 2025, and cites a combined Working Families Party and Democratic Socialists of America membership above 700,000 as a political base. A recurring argument across essays is that a future administration should confront judicial limits on agency authority directly rather than attempt to draft rules that survive them.
Roosevelt is a progressive organization, affiliated with the Franklin D. Roosevelt Presidential Library, and is foundation-supported, with the Ford and Hewlett foundations among its reported funders. The compendium is best read as a statement of programmatic intent rather than empirical analysis.
7. CAP and EPI: two progressive analyses of immigration enforcement costs
The Center for American Progress published an analysis on August 4 by Ben Greenho examining a May 21, 2026 U.S. Citizenship and Immigration Services memorandum directing most green-card-eligible applicants toward consular processing abroad rather than domestic adjustment of status. CAP’s figures indicate adjustment of status produced more than 740,000 green cards in FY2025, including 421,820 for spouses, children, and parents of U.S. citizens, and its modeling suggests routing FY2022–25 family-based cases overseas would have increased consular processing volume by 49 to 75 percent. The analysis notes that visa interview waits already exceeded a year in some countries as of June 2026 and that immigrant visa processing is suspended in 75 countries. CAP is a progressive advocacy organization founded in 2003 by John Podesta and funded by foundations, corporations, labor, and individual donors; the estimates are its own projections rather than agency data, and the government’s stated rationale is set out in the memorandum itself.
Separately, the Economic Policy Institute released technical documentation for a “Cost of Deportations” calculator on August 5, produced by Monique Morrissey with the Institute for Policy Studies’ National Priorities Project and partners. The tool allocates a projected $268.9 billion in federal immigration enforcement spending — an average of $2,358 per taxpayer, by EPI’s calculation — down to state, county, and congressional district level using IRS Statistics of Income tax-share data, then converts each jurisdiction’s share into equivalents such as teacher or nurse salaries. EPI is a progressive institution and has historically drawn a substantial share of its funding from organized labor. The methodology is documented and reproducible, which is to its credit; the framing of the trade-offs is an editorial choice readers should recognize as such.
8. Niskanen Center: surprise-billing arbitration keeps growing
The Niskanen Center published New Data, Same Problem: No Surprises Act Arbitration Abuse Persists on July 30, analyzing newly released CMS data covering all of calendar 2025 — the first complete-year picture of the independent dispute resolution system.
The numbers are striking against the original projection. Disputes rose roughly 75 percent year-over-year to more than 2.5 million initiated and about 2.2 million decided, against a federal estimate of approximately 17,000 cases per year. Median winning provider offers ran about four times the statutory qualifying payment amount across the year and five times in the fourth quarter, reaching roughly 30 times QPA in neurology and 15 times in surgery. Total program cost exceeded $2.8 billion, including more than $1.2 billion paid to arbitrators in 2025 alone. Four filers — HaloMD at 19 percent, TeamHealth at 12 percent, SCP Health at 10 percent, and Radiology Partners at 7 percent — accounted for nearly half of all disputes. Niskanen recommends replacing the arbitration mechanism with rate benchmarking, or amending the statute to bar arbitrators from weighing historical rates.
Niskanen occupies a center-out, reformist position and is foundation-funded, with Arnold Ventures and the Hewlett Foundation among reported supporters. Its analysis rests on the government’s own published data, which makes it among the more directly verifiable claims surveyed this week.
Also relevant to TIJ beats
Three further releases bear on this publication’s accountability coverage. The R Street Institute published its first Election Policy 50-State Scorecard on July 28, grading every state across six categories on statutes in effect as of June 15, 2026; the average rating was four stars, every state scored at least three, and Florida, Georgia, Ohio, and Utah earned five. A non-scored supplemental section covering primaries, political speech, and election-worker safety averaged just 2.5 stars — R Street’s implicit finding being that states have largely fixed election mechanics while neglecting the surrounding institutional infrastructure. R Street is a center-right free-market organization funded by foundations and corporations, which it discloses.
On fiscal policy, the Tax Foundation released fifteen capital-investment case studies in late July, modeling projects across energy, manufacturing, and technology under five cost-recovery regimes. Alex Muresianu’s analysis finds that moving from MACRS depreciation with R&D amortization to full expensing raises average internal rate of return by 1.57 percentage points, with the One Big Beautiful Bill Act’s provisions alone accounting for 0.88 points; gains concentrate where 39-year structures dominate upfront cost, and the paper recommends extending expensing to structures. The Tax Foundation, founded in 1937, is business-oriented and generally center-right, funded by individuals, foundations, and corporations.
Finally, Brookings published a detailed explainer on August 4 by Sarah Reber and Sarah Turner tracing income-driven student loan repayment from the 1994 ICR program through the litigation-halted SAVE plan to the new Repayment Assistance Plan. The authors argue RAP could meaningfully simplify a fragmented system but caution that implementation will strain a significantly downsized Federal Student Aid office. Brookings is centrist to center-left and discloses funding from foundations, corporations, individuals, and some foreign governments. RAND separately published a design study for secure AI inference data centers on August 4, estimating $37 million to $50 million for a proof-of-concept facility and $277 million to $345 million at enterprise scale, buildable in as few as 14 months with off-the-shelf hardware — a concrete cost figure for a debate usually conducted in abstractions.
All figures above are drawn from the linked publications and reflect each institution’s own reported data and methodology. The Investigative Journal did not independently replicate the underlying analyses. Where an organization is an advocacy institution, its estimates are identified as projections rather than official data. Institutions named in this roundup may respond to editor@tij.news.