HB 1032 died; Indiana redistricting turns on AI evidence

HB 1032 died in the Indiana Senate, 31-19, on December 11, 2025. That fact matters first because it fixes the current legal baseline: as of Q3 2026, Indiana’s 2021 congressional districts remain in place for the 2026 cycle, and no Indiana-specific lawsuit over a newly enacted mid-decade congressional map has been confirmed in the materials reviewed here.[1][2]

It matters second because the bill was not just a map bill. For lawyers tracking Indiana redistricting and the role of algorithmic evidence, the question now begins with a failed enactment that previewed both the mapmaking theory and the litigation architecture likely to return. The bill page said HB 1032 would have allowed the General Assembly to amend congressional districts and specified requirements for actions challenging the map.[3] Reported Section 3 language went further: it would have barred temporary restraining orders and routed appellate review exclusively to the Indiana Supreme Court.[4]

That is not a small drafting flourish. In election litigation, emergency relief is often the difference between a live remedy and an academic opinion issued after ballots, filing deadlines, or primaries have already moved past the dispute. A statutory ban on temporary restraining orders would have forced challengers into a narrower timing channel before any court reached the map’s merits. Exclusive appellate jurisdiction in the Indiana Supreme Court would have narrowed the institutional path as well. Rep. Matt Pierce reportedly said he had never seen such provisions in a General Assembly bill.[4]

So the cleanest verification frame is this: HB 1032 failed; the 2021 map remains operative for 2026; the bill’s procedural clauses are not live Indiana law; and the legal-risk question is anticipatory, not a substitute for advice about any filed case. The useful work is to read the bill as a template.

What HB 1032 put on the table

The Senate defeat was not close procedurally, even if it was politically awkward: 21 of 40 Republicans joined all 10 Democrats in voting no.[1] But the failed bill still supplies the record any renewed effort will have to work around.

Smaltz’s reported testimony is unusually direct for later litigation because it separated the map’s purpose from any invented neutrality story. He said the map was drawn by the National Republican Redistricting Trust using recent-election data and that districts were “specifically drawn for a political advantage.”[4] That testimony does not make the bill law, and it does not by itself decide any constitutional claim. It does, however, make it harder for a future defense to pretend that partisan objectives were incidental if a substantially similar map returns.

The procedural provisions would matter just as much. A future plaintiff would not merely be asking whether the map diluted votes, violated state constitutional guarantees, or crossed a federal racial-gerrymandering line. The plaintiff might first have to fight about whether emergency relief is available at all, whether trial-court review can meaningfully pause election administration, and how quickly appellate review concentrates in one court.

The map mechanics are concrete enough to matter, but not enough to decide the case

The proposed map’s politics were not abstract. House Democratic Leader Phil GiaQuinta said the plan split Marion County four ways and that the 7th District would be 31% Black and 16% Hispanic.[5] Those figures are the sort of facts litigators would pin to a complaint because they identify where representational injury might be concentrated.

Still, those mechanics do not automatically supply a winning federal claim. Marion County’s split could support a story about community fracture. The 7th District’s racial composition could become relevant to vote-dilution or racial-gerrymandering analysis. But after Louisiana v. Callais, a challenger cannot simply point to racial impact and partisan motive, attach a cleaner-looking alternative map, and expect federal Section 2 doctrine to do the rest.

The question has shifted. It is no longer only whether HB 1032 was partisan. Smaltz’s reported testimony largely answers that at the political-purpose level. The harder federal question is what kind of illustrative maps can survive evidentiary scrutiny when partisanship is treated as a legitimate state objective that the plaintiff’s alternatives must also satisfy.

Callais narrows the federal route

Louisiana v. Callais is binding federal doctrine, not just another datapoint in the redistricting calendar. Decided 6-3 on April 29, 2026, with Justice Alito writing, Callais addressed the “increased use and capabilities of computers in drawing districts and creating illustrative maps” and tightened what plaintiffs must show when they use illustrative maps in a Section 2 case.[6]

For an Indiana challenge, the important part is operational. A federal plaintiff’s illustrative maps would need to satisfy every legitimate state objective, including partisan objectives; the maps could not use race as a districting criterion; and racial-bloc-voting analysis would need to control for party affiliation.[6] A more compact explanation of those doctrinal changes is available in What the Supreme Court Changed About Section 2 in Louisiana v. Callais.

That framework creates an uncomfortable evidentiary assignment for challengers. If the state says it pursued partisan advantage, the plaintiff cannot answer only with an alternative map that improves minority electoral opportunity while ignoring partisan performance. The plaintiff has to generate maps that respect the state’s legitimate criteria as the Court now defines them, then show that the enacted map is still legally defective when compared with that constrained universe.

This is where “AI” is too broad a label and “algorithmic illustrative maps” is closer to the actual litigation problem. The fight is not whether software drew a pretty boundary. It is whether the simulation rules are declared, reproducible, and legally matched to the state’s objectives. Which elections were used to measure partisan performance? Were recent elections weighted equally or selectively? Were incumbent-protection assumptions included? Did the map generator exclude race entirely, or did race enter through seed selection, constraints, tuning, or post-processing? Did the racial-bloc analysis separate race from party well enough to satisfy Callais?

A plaintiff who cannot answer those questions will have a presentation problem before reaching the merits. A defense team that can show the plaintiff’s simulations quietly disfavored the state’s partisan objective may not need to prove the enacted map is admirable. It may only need to show that the comparator set is legally mismatched.

Why the NRRT record points directly into the Callais problem

The Indiana record already contains the fact pattern Callais makes awkward. The reported testimony was that NRRT used recent-election data to draw for political advantage.[4] That means a challenger’s federal illustrative maps would likely have to model that partisan objective rather than treat it as contamination to be removed.

Texas shows why that distinction matters. In Texas litigation involving NRRT-linked testimony, Adam Kincaid reportedly denied using racial data; a three-judge panel later ruled the challenged map an unconstitutional racial gerrymander, and the Supreme Court stayed that ruling on December 4, 2025.[7] The Texas posture is not an Indiana holding. It is useful because it shows how parties now litigate the boundary between partisan data, racial data, and the inferences courts may draw from map effects.

Indiana challengers would have to be ready for the same evidentiary pivot. If the state defends a map as partisan, the plaintiff’s burden under Callais is not satisfied by arguing that partisan intent is distasteful. The plaintiff would need to show, through compliant illustrative maps and supporting analysis, that the challenged districts fail federal law even after the state’s partisan objective is built into the comparator universe.

Simulation evidence can decide cases, but assumptions can decide simulations

Courts have already shown that simulation evidence can carry real weight. In Utah, Judge Dianna Gibson struck down Map C after finding it “more Republican than over 99% of expected maps,” a statistic reported through coverage of the order.[8] That was not a decorative chart at the back of a brief. It was a comparison that helped define the legal significance of the enacted map.

But simulations do not become neutral merely because they are computational. Chen and Rodden’s Florida work is the cautionary example. Changing the election baseline from Bush-Gore 2000 to Obama-McCain 2008 reversed the conclusion: the modal GOP-leaning outcome moved from 16 of 25 districts to 13 of 27 districts.[9] That is not a minor sensitivity check. It shows how a model can appear empirical while carrying a decisive legal or political assumption inside its inputs.

For Indiana, that means the first fight over algorithmic evidence would likely be a fight over design choices. A challenger using recent statewide elections would have to justify which elections count as recent, whether federal and state contests are comparable, and how to handle turnout variation. If the state says HB 1032’s successor pursued a specific partisan performance target, the plaintiff has to decide whether to include that target as a constraint, a scoring variable, or an output measure. Each choice changes what the simulated universe means.

A defensible illustrative-map file would therefore need more than boundary shapefiles. It would need a record of constraints, seed rules, election inputs, excluded variables, compactness measures if used, population tolerances, treatment of county and municipal splits, and the exact point at which race was excluded from the map-generation process. If race appears only later, in legally required performance analysis, that sequence has to be documented. If party is controlled in racial-bloc analysis, the method has to be explainable to a judge who is not buying software but deciding admissibility and weight.

This is also where legal-tech buyers should be careful about vendor promises. A tool that can generate thousands of maps is not automatically a Callais-ready tool. The relevant question is whether the tool can preserve the assumptions a lawyer must disclose and defend: election baselines, legitimate state objectives, race exclusions, partisan constraints, audit trails, and reproducibility. Volume is not the same thing as admissible comparison.

The state-law route remains open, but not settled

A federal Section 2 claim is not the only conceivable route. Rep. Pierce identified a possible claim under Article 2, Section 1 of the Indiana Constitution—the state’s free-and-equal-elections provision—as a case of first impression.[4] That makes the theory worth tracking and unsafe to oversell.

The attraction is obvious. A state constitutional claim would not have to pass through Callais in the same way a federal Section 2 claim would. Indiana courts could decide what “free and equal” means for congressional redistricting under state law, including whether extreme partisan line-drawing, procedural foreclosure, or fractured political communities violate the state constitution.

The problem is equally obvious: the materials reviewed do not identify an Indiana redistricting standard already supplied by state courts under Article 2, Section 1. First impression means counsel would be building the doctrine while asking for relief. That can be an advantage when federal doctrine has narrowed, but it also increases uncertainty over threshold questions: justiciability, manageable standards, timing, remedy, and the degree of deference owed to the General Assembly.

The failed HB 1032 litigation provisions would have sharpened those state-law questions. A temporary-restraining-order ban would not merely affect federal claims; it would also shape how quickly a state constitutional plaintiff could stop a map from governing an election. Exclusive Indiana Supreme Court appellate jurisdiction would place the development of any first-impression standard on a compressed and centralized path.

Kagan’s dissent captures the stakes, but not the whole task

Justice Kagan’s Callais dissent argued that the decision made Section 2 “all but a dead letter” because a state could insulate a map by announcing a partisan gerrymander.[10] For Indiana, that criticism lands with force because HB 1032’s lead author reportedly described the map as drawn for political advantage.[4]

But a litigation memorandum cannot stop at the dissent. The controlling opinion sets the burden for federal court. If a renewed Indiana plan is challenged under Section 2, the plaintiff’s evidence will be tested against Callais’s requirements, not against the dissent’s warning about how hard those requirements are to satisfy.

That is the practical meaning of the post-Callais evidentiary landscape. A challenger must be prepared to prove that its illustrative maps are legally comparable even when partisan objectives are included, that race did not drive the map generation, and that racial bloc voting has been analyzed with party controlled. The politics of HB 1032 may explain why a plaintiff sues. The simulations may decide whether the plaintiff survives.

Why recurrence is a legal-risk issue, not a 2028 prediction exercise

Sabato’s Crystal Ball has forecast that Indiana could return to the redistricting fight in 2028.[11] That is an analyst projection, not a legal obligation and not evidence that a particular bill will be filed. Its value is narrower: it explains why a dead 2025 bill should be preserved in research files rather than discarded as old legislative theater.

If a successor bill borrows HB 1032’s structure, counsel will need to look at three things immediately: whether the procedural restrictions return, whether the map again relies on recent-election data for partisan advantage, and whether any affected district creates a plausible federal or state-law plaintiff. The order matters. Emergency-relief limits can shrink the window for factual development before anyone has finished arguing about the map’s legality.

• For a state-law challenge, the unresolved question is whether Indiana courts will recognize a manageable Article 2, Section 1 standard for congressional redistricting.
• For a federal Section 2 challenge, the immediate question is whether plaintiffs can build Callais-compliant illustrative maps that include legitimate partisan objectives while excluding race from map generation.
• For tool selection, the question is whether simulation software can preserve assumptions, audit trails, and reproducible outputs rather than merely generate many alternatives.

Indiana’s 2026 congressional map is unchanged. The harder lesson from HB 1032 is that the next mid-decade attempt will not be judged mainly by slogans about partisan advantage. The case will turn on timing rules, available remedies, and the design, assumptions, and admissibility of algorithmically generated illustrative maps, with the state-law route still open and the federal route sharply narrowed.

References

1. Indiana Senate rejects congressional redistricting bill, WFYI, December 11, 2025.
2. Indiana congressional redistricting status, National Conference of State Legislatures.
3. HB 1032, Indiana General Assembly.
4. The Indiana Citizen redistricting coverage, The Indiana Citizen, December 8, 2025.
5. The Indiana Citizen map mechanics coverage, The Indiana Citizen, May 1, 2026.
6. Louisiana v. Callais, Supreme Court of the United States, April 29, 2026.
7. Texas redistricting testimony and litigation coverage, Democracy Docket, October 2025.
8. Utah redistricting order coverage, Democracy Docket.
9. Unintentional Gerrymandering: Political Geography and Electoral Bias in Legislatures, University of Chicago Legal Forum.
10. Supreme Court decision in Louisiana redistricting case, SCOTUSblog, April 29, 2026.
11. Sabato’s Crystal Ball redistricting forecast, Sabato’s Crystal Ball.