Provenance · The Debate
The debate behind Trump’s Deportation Machine Is Being Built Through Budgets, Beds and Deputies
The questionMass Deportation Requires a Shadow Infrastructure — and Trump Is Building It in Plain Sight
How this debate works
Before writing, The Arbiter stress-tests each story by framing the two strongest opposing positions and arguing both sides of a structured three-round debate: opening arguments, rebuttals, then steel-manning the opponent and answering one question — what specific, verifiable evidence would change my mind?
Both advocacy roles were generated by the same provider/model: Advocate A · OpenAI GPT-5.5 argued the first position, and Advocate B · OpenAI GPT-5.5 argued the second. The published article was written separately by OpenAI GPT-5.5: the debate supplies the questions, and the author verifies key claims with its own research before taking a position.
Sources in this transcript are evidence as each advocate presented it during the debate — research leads, not independently verified endorsements.
Why we covered this
Publish. Moves immigration coverage away from rhetoric and raids toward the operational machinery: detention capacity, oversight erosion and legal exceptionalism. High policy relevance and enough institutional detail for a strong analysis.
The positions
Advocate A · OpenAI GPT-5.5 argued
Trump’s mass-deportation agenda is not merely campaign rhetoric but an institutional project: by expanding detention capacity, outsourcing enforcement, narrowing due-process protections and attacking oversight, the administration is building a semi-permanent deportation apparatus that could outpace courts and democratic checks before they can respond.
Advocate B · OpenAI GPT-5.5 argued
The alarm over a “shadow infrastructure” overstates what is actually happening: expanded detention, faster removals and tougher enforcement are lawful tools for restoring control to a broken immigration system, and the real constraint is not civil society resistance but whether the federal government can finally execute laws that Congress has left underenforced for decades.
Round 1 — Opening arguments
Each advocate states its position and strongest evidence.
- The January 20, 2025 executive order directed DHS to expand detention capacity through resources and contracts, use expedited removal authorities, and broaden 287(g) state-local enforcement agreements.
- The July 2025 reconciliation law supplied roughly $170 billion for immigration enforcement and border security, including $45 billion for ICE detention and about $30 billion for ICE hiring.
- ICE’s 287(g) agreements reportedly increased from 135 in December 2024 to more than 1,300 across forty states by January 2026.
- The administration expanded faster-removal mechanisms while the immigration-court backlog was about 3.8 million cases and the number of immigration judges was capped at 800.
- Oversight capacity was weakened through reported elimination of immigration oversight offices and denial of congressional access to detention facilities.
Read the full argument
My position is that the mass-deportation agenda should be understood less as episodic campaign theater and more as institutional construction: a durable enforcement system that expands detention supply, recruits non-federal actors, accelerates removals, and reduces the practical capacity of courts and overseers to intervene in time. The strongest evidence is not rhetoric; it is budgeting, contracting, delegation, and procedure.
First, the administration has moved the bottleneck from “who can be arrested?” to “how many people can the system hold and process?” Executive Order 14159, issued January 20, 2025, directed DHS to allocate legally available resources or establish contracts to “construct, operate, control, or use facilities” to detain removable noncitizens, and to pursue efficient and expedited removals. That was followed by appropriations: CFR reports that the July 2025 reconciliation law provided about $170 billion for immigration enforcement and border security over four years, including $45 billion for ICE detention and roughly $30 billion for ICE hiring. The American Immigration Council estimates that this funding could allow ICE to acquire capacity for about 135,000 detainees at a time, more than triple the system’s capacity when Trump took office. Even if that estimate is contested, the order of magnitude matters: detention capacity is the indispensable infrastructure of mass deportation.
Second, the project relies on outsourcing. The same executive order directs DHS to use 287(g) agreements so state and local officers can perform immigration-officer functions. CFR reports that ICE had more than 1,300 such agreements across forty states by January 2026, compared with 135 in December 2024. That is not simply “more ICE agents”; it is a force-multiplier model that embeds federal immigration enforcement into ordinary local policing, jails, and traffic stops. This makes enforcement geographically diffuse and harder for any single court order, state policy, or local protest to constrain.
Third, the due-process and oversight channels are being narrowed as volume rises. Expedited removal allows deportation without a full immigration-court hearing; the administration has expanded its use while the immigration-court backlog was about 3.8 million cases as of June 2025, according to CFR. Meanwhile, the reconciliation law capped immigration judges at 800 despite increased enforcement volume. The Brennan Center also reports that Congress gave ICE $45 billion for detention while ICE denied lawmakers access to facilities despite statutory oversight rights, and that DHS eliminated oversight offices including the Office of the Immigration Detention Ombudsman and the Office for Civil Rights and Civil Liberties.
The steel-man defense is that immigration enforcement is legally authorized, Congress appropriated funds, and a government may prioritize removal of people unlawfully present. But the debate here is institutional risk. When detention beds, deputized local officers, expedited procedures, and weakened oversight expand together, the system can generate removals faster than individualized legal claims, habeas petitions, journalists, legislators, or civil-society monitors can detect and correct errors. That is what is at stake: not merely stricter enforcement, but whether the enforcement machine becomes semi-permanent before democratic checks can meaningfully operate.
Evidence cited by Advocate A · OpenAI GPT-5.5 (4) — as presented, not independently endorsed
- Executive Order 14159 / govinfo PDFSupports that the administration formally ordered expanded detention facilities, expedited removals, and 287(g) state-local enforcement agreements.
- Council on Foreign Relations, “How Trump Is Reshaping Immigration Enforcement”Supports claims about the July 2025 funding package, immigration-court backlog and judge cap, expanded expedited removal, and growth of 287(g) agreements.
- American Immigration Council, “Immigration Detention Expansion in Trump’s Second Term”Supports the estimate that new detention funding could enable ICE capacity of roughly 135,000 people at a time and that oversight has become more opaque.
- Brennan Center for Justice, “Budget Bill Massively Increases Funding for Immigration Detention”Supports claims about the scale of detention funding, private-facility expansion, denial of congressional access, and elimination of oversight offices.
- 8 U.S.C. § 1357(g) expressly authorizes trained state and local officers operating under written agreements to perform immigration-officer functions subject to federal law and consistency with state and local law.
- Expedited removal under 8 U.S.C. § 1225(b)(1) and post-final-order removal under 8 U.S.C. § 1231 are statutory mechanisms created by Congress, not extra-legal innovations.
- ICE reported more than 6.1 million people on its non-detained docket in FY2024, and official data showed about 1.45 million non-detained people with final orders of removal as of November 24, 2024.
- Historical removal levels were already high before Trump’s second term, with ICE reporting 368,644 removals in FY2013 including 101,000 expedited-removal orders.
Read the full argument
Advocate A’s weakest point is the inference that scale plus speed equals a “shadow infrastructure” or “legal exceptionalism.” That framing treats the ordinary instruments of immigration enforcement—detention beds, 287(g) agreements, expedited removal, and appropriated funds—as suspect because they are effective. But the key empirical question is whether these tools are extra-legal or whether they implement statutory commands Congress already enacted. The record supports the latter. Section 287(g) is not an improvisation: 8 U.S.C. § 1357(g) expressly allows written agreements with state or local officers to perform immigration-officer functions, requires training and adherence to federal law, and limits performance to what is consistent with state and local law. Expedited removal is likewise a congressional creation in 8 U.S.C. § 1225(b)(1), adopted in the 1996 IIRIRA framework, not a novel workaround of courts. Post-order detention and removal are also statutory: 8 U.S.C. § 1231 directs removal within a removal period after a final order, with detention authority for specified categories. A may argue these tools are dangerous when expanded, but “dangerous when expanded” is different from “shadow” or “exceptional.”
A’s framing is also incomplete because it minimizes the enforcement deficit the system has accumulated. ICE’s FY2024 annual report says its non-detained docket exceeded 6.1 million and grew by about 24.6% that year, while ICE removed 271,484 people in FY2024. Separately, official ICE data released in early 2025 showed roughly 1,445,549 people on the non-detained docket already had final orders of removal as of November 24, 2024. Those are not merely pending asylum seekers awaiting adjudication; they are cases where the legal process has ended. If the government lacks detention, transportation, personnel, and local cooperation to execute final orders, then the practical message is that court orders are optional. The normative claim I am making is explicit: a sovereign legal system has a legitimate interest in making final judgments enforceable.
A’s own evidence also undercuts the idea that courts and democratic checks are being bypassed. CFR reports that a federal judge in February 2026 ruled the administration’s third-country deportation policy unlawful on statutory and due-process grounds. That is not proof of a machine beyond constraint; it is proof that judicial review remains capable of stopping unlawful applications. Likewise, the July 2025 funding package was enacted by Congress. Critics may dislike the appropriation, but public statutory funding is the opposite of clandestine construction.
The strongest contrary evidence is historical continuity. ICE conducted 368,644 removals in FY2013, including 101,000 after expedited-removal orders; FY2025 removals reportedly reached 442,637. The difference is degree and prioritization, not the invention of an underground apparatus. The real constraint is administrative capacity to enforce laws Congress already put on the books, not civil society’s ability to delegitimize enforcement by labeling lawful capacity “shadow infrastructure.”
Evidence cited by Advocate B · OpenAI GPT-5.5 (5) — as presented, not independently endorsed
- Legal Information Institute, 8 U.S.C. § 1357Supports that Congress authorized 287(g) agreements with trained state and local officers to perform immigration functions under written agreements.
- Legal Information Institute, 8 U.S.C. § 1225Supports that expedited removal is a statutory process codified in federal immigration law.
- Legal Information Institute, 8 U.S.C. § 1231Supports that federal law provides for detention and removal after a final order of removal.
- ICE FY2024 Annual ReportSupports FY2024 removal totals and the size and growth of ICE’s non-detained docket.
- ICE FY2013 Immigration RemovalsSupports that ICE conducted 368,644 removals in FY2013 and 101,000 removals after expedited-removal orders.
Round 2 — Rebuttals
Each advocate concedes its opponent’s best point, then explains why its position still holds.
- The strongest pro-B point is that 287(g), expedited removal, and detention after final removal orders are statutory mechanisms, not newly invented illegal tools.
- The administration’s expansion of expedited removal matters because it can bypass immigration-court hearings for people inside the United States, and constitutional challenges to that interior use remain unresolved.
- ICE’s July 2025 funding and detention expansion represent durable institutional capacity, including beds, contracts, staffing, and transport infrastructure, not merely short-term campaign signaling.
- Court rulings do not fully constrain the apparatus if removals, transfers, and detention harms occur before judicial review can provide individualized relief.
- Oversight weakening is central because an enforcement system can be publicly funded yet function as a shadow infrastructure if legislators, ombuds offices, and civil-rights monitors cannot inspect it effectively.
Read the full argument
B’s strongest point has genuine merit: 287(g), expedited removal, detention after final orders, and appropriated enforcement funding are not inherently extra-legal. They are statutory tools, and the existence of roughly 1.45 million non-detained people with final removal orders is a real administrative problem for any government committed to enforcing immigration judgments. I do not need to deny either point. The dispute is whether lawful tools, when scaled together and paired with reduced oversight and compressed procedure, become an institutional apparatus that can outpace meaningful checks. On that question, A’s position still holds.
The crucial distinction is between formal legality and operational accountability. CRS explains that the January 2025 executive order called for expedited removal “to the fullest extent authorized by Congress,” and that people subject to expedited removal bypass ordinary administrative removal proceedings and the opportunity to present their cases to immigration judges. CRS also notes that the expansion potentially covers people found anywhere in the United States who cannot prove two years’ presence, and that constitutional challenges remain unresolved because earlier appellate review addressed statutory consistency rather than the constitutional sufficiency of applying expedited procedures to people inside the country. So B is right that Congress created expedited removal; but A is right that expanding it inside the country changes the practical due-process environment.
Nor does congressional appropriation disprove institutional entrenchment. Brennan reports that ICE received a $75 billion cash infusion through July 2025 funding, including $45 billion for new detention centers, fully funding ICE until 2029 and more than tripling its annual budget. It also reports that detention rose from fewer than 40,000 people when Trump took office to about 70,000, with plans to add more than 100,000 beds by November 30. That is not merely ordinary enforcement catching up to final orders; it is a durable capacity buildout that creates contracts, facilities, staffing, and transport infrastructure that future officials may find difficult to unwind.
B’s court-check argument is also incomplete. CFR’s example of a February 2026 ruling against third-country deportations shows courts still exist, but it also shows the administration had already spent about $40 million sending hundreds of migrants to third countries before the ruling. Judicial review after-the-fact is not equivalent to timely individualized process before removal, especially when detention, transport, and removal capacity expand faster than representation and court capacity.
Finally, the oversight evidence directly answers B’s “public statutory funding” point. Brennan reports that DHS effectively eliminated key oversight offices and that ICE denied lawmakers access to facilities despite statutory access rights. Public funding can still produce a shadow infrastructure if the machinery is visible but its conditions, error rates, and individual cases are inaccessible to the institutions charged with monitoring it.
Evidence cited by Advocate A · OpenAI GPT-5.5 (4) — as presented, not independently endorsed
- Congressional Research Service, “Recent Executive Branch Actions on Immigration (Part 2)”Supports that the January 2025 executive order called for expanded expedited removal, detention facilities, and 287(g) agreements, and that expedited removal bypasses ordinary immigration-judge proceedings while constitutional challenges remain pending.
- Brennan Center for Justice, “How ICE’s Budget Boom Is Changing Immigration Detention”Supports claims about ICE’s $75 billion funding infusion, $45 billion for new detention centers, detention rising to about 70,000 people, planned bed expansion, ICE-owned aircraft, and diminished oversight.
- Council on Foreign Relations, “How Trump Is Reshaping Immigration Enforcement”Supports claims about the expansion of ICE interior operations, growth of 287(g) agreements to more than 1,300 across forty states, and the February 2026 ruling against third-country deportations after prior removals had occurred.
- Federal Register, Executive Order 14159, “Protecting the American People Against Invasion”Supports that the administration formally made total and efficient enforcement, detention capabilities, and execution of final removal orders central federal policy.
- A conceded that 287(g), expedited removal, detention after final orders, and appropriated enforcement funding are statutory tools rather than newly invented illegal mechanisms.
- 8 U.S.C. § 1225(b)(1) and 8 U.S.C. § 1357(g) show that Congress authorized expedited removal and state-local immigration enforcement agreements subject to specified conditions.
- ICE reported that its non-detained docket grew from more than 6.1 million to more than 7.6 million in FY2024 while ICE removals totaled 271,484 that year.
- Officially reported data stated that about 1.45 million people on ICE’s non-detained docket had final removal orders as of November 24, 2024.
- Oversight problems justify inspection, litigation, and statutory correction, but they do not by themselves prove that lawful enforcement capacity is a shadow or extra-legal infrastructure.
Read the full argument
A’s rebuttal concedes the two propositions that most weaken the “shadow infrastructure” thesis: the tools at issue are statutory, and the government faces a large stock of final removal orders that have not been executed. That concession matters because it shifts the dispute from whether the administration is building an extra-legal machine to whether lawful enforcement capacity should be expanded when existing capacity is inadequate. If the law authorizes detention, expedited removal, post-order removal, and state-local cooperation, then A must show not merely that scale creates risk, but that scale itself defeats legality. A has not done that.
The core weakness is that A treats “compressed procedure” as synonymous with insufficient procedure, even where Congress deliberately chose faster procedure for defined categories. Section 1225(b)(1) authorizes expedited removal for certain inadmissible noncitizens and provides credible-fear screening and limited immigration-judge review in asylum-related cases; Section 1357(g) authorizes state-local agreements under federal direction, training, and limits consistent with state and local law. These are not loopholes around Congress; they are Congress’s enforcement design. A can reasonably argue for more process as a normative preference, but that is not the same as proving “legal exceptionalism.”
A’s final-order concession is even more consequential. ICE’s FY2024 report says the non-detained docket grew from more than 6.1 million to more than 7.6 million, while ICE removed 271,484 people in FY2024. Officially reported data also stated that about 1.45 million people on the non-detained docket had final removal orders as of November 24, 2024. If final removal orders are routinely unenforced, then the binding output of immigration courts becomes practically contingent on ICE’s logistical capacity. Detention beds, transport, and local cooperation are therefore not inherently suspicious infrastructure; they are the means by which final judgments become real.
A also overreads oversight disputes. ICE’s own public materials describe detention facilities as subject to national detention standards, inspections, audits, DHS Inspector General review, ICE Office of Detention Oversight review, and other compliance mechanisms. Those mechanisms may be imperfect, and allegations of access denial should be investigated. But imperfection supports targeted oversight remedies, not the conclusion that expanded detention is “shadow” governance. A’s cited court example also cuts both ways: if courts can enjoin unlawful applications, the system is not beyond legal constraint.
The strongest remaining B position is therefore limited but decisive: expanded capacity can be abused, but abuse is not shown by capacity itself. A’s argument depends on converting lawful scale into illegitimacy without proving that the relevant statutory checks, judicial review, and inspection mechanisms have been nullified. The core issue is whether the federal government can execute immigration laws Congress already enacted; A’s rebuttal acknowledges that legal foundation and never explains how a government should enforce 1.45 million final orders without materially expanding operational capacity.
Evidence cited by Advocate B · OpenAI GPT-5.5 (5) — as presented, not independently endorsed
- Legal Information Institute, 8 U.S.C. § 1225Supports that expedited removal is codified by Congress and includes statutory procedures such as credible-fear screening and limited immigration-judge review.
- Legal Information Institute, 8 U.S.C. § 1357Supports that Congress authorized 287(g) agreements allowing trained state and local officers to perform immigration functions under written agreements and federal-law constraints.
- ICE FY2024 Annual ReportSupports FY2024 removal totals and the reported growth of ICE’s non-detained docket from more than 6.1 million to more than 7.6 million.
- Federal Register public inspection notice citing ICE final-order dataSupports the reported figure that 1,445,549 people on ICE’s non-detained docket had final removal orders as of November 24, 2024.
- ICE Facility Inspections pageSupports that ICE detention facilities are subject to detention standards and multiple inspection or audit mechanisms, including DHS OIG and ICE Office of Detention Oversight review.
Round 3 — Steel man & the crux
Each advocate argues the strongest version of the opposing case before answering: what evidence would change my mind?
- The strongest pro-B point is that the main enforcement tools are statutory and that many non-detained people have final removal orders, so some capacity expansion is legally and administratively defensible.
- A’s stronger claim is institutional rather than facial-legality based: lawful tools can still create an accountability deficit when scaled together faster than courts, counsel, inspectors, and Congress can review individual cases.
- Expanded interior use of expedited removal is especially important because it can bypass ordinary immigration-court hearings for people unable to prove two years of continuous presence in the United States.
- The current enforcement buildout is not limited to executing final removal orders; it also expands detention, local deputization, raids, and expedited-processing capacity across the system.
- Judicial review remains real but may be too slow to prevent erroneous removals or detention harms when enforcement operations proceed before facts are visible to courts or overseers.
Read the full argument
B’s strongest argument is that the instruments A criticizes are not invented outside the law: Congress authorized expedited removal, 287(g) agreements, detention after final orders, and then appropriated money for enforcement. B can also fairly say that roughly 1.45 million non-detained people with final removal orders create a serious legitimacy problem for the immigration courts if final judgments are routinely unenforced.
That is the best version of B’s case, and it narrows the dispute appropriately. My claim is not that every component is facially illegal. It is that legality at the level of isolated statutory tools does not settle whether the combined institutional design can overwhelm the checks that make legality meaningful in individual cases. A scalpel and a factory assembly line can both be lawful tools; the institutional question is what happens when faster removal authority, vastly larger detention capacity, deputized local enforcement, and diminished oversight are expanded simultaneously.
The specific reason A remains stronger is that B treats statutory authorization as if it automatically supplies operational accountability. It does not. Expedited removal under 8 U.S.C. § 1225 can result in removal “without further hearing or review” except for narrow asylum-related and status-claim review channels. CRS explains that the January 2025 policy potentially applies that mechanism to people found anywhere in the United States who cannot prove two years of continuous presence, while constitutional challenges to that interior application remain unresolved. That matters because the risk is not only wrongful legal theory; it is wrongful factual sorting at speed, especially for people with limited documents, language access, or counsel.
B’s final-order argument is strongest when enforcement targets people whose cases have fully concluded. But the infrastructure being built is not limited to that subset. CFR reports an enforcement model that includes expanded expedited removal, nationwide raids, more than 1,300 287(g) agreements by January 2026, and nearly $170 billion in new enforcement and border-security funding, including $45 billion for ICE detention. Brennan reports detention rising to about 70,000 people and plans for major bed expansion, while also describing weakened oversight access. That combination supports A’s thesis: the apparatus is designed not merely to execute settled court orders, but to increase arrests, detention, rapid processing, and removals at a systemwide scale.
B is right that courts still can intervene. But delayed court intervention is an incomplete check when removal, third-country transfer, or prolonged detention can occur before counsel, legislators, inspectors, or judges obtain facts. The question is comparative speed: whether enforcement capacity grows faster than individualized review and oversight. On the present record, the documented expansion of beds, money, deputized actors, and expedited procedures is more concrete than B’s assurance that formal statutory existence will keep the apparatus democratically accountable.
Evidence cited by Advocate A · OpenAI GPT-5.5 (5) — as presented, not independently endorsed
- Legal Information Institute, 8 U.S.C. § 1225Supports B’s point that expedited removal is statutory, and A’s point that the statute permits removal without further hearing or review in specified circumstances subject to limited exceptions.
- Legal Information Institute, 8 U.S.C. § 1357Supports B’s point that 287(g) state-local immigration enforcement agreements are authorized by Congress under federal supervision and written agreements.
- Congressional Research Service, “Recent Executive Branch Actions on Immigration (Part 2)”Supports that the January 2025 executive order directed expansion of expedited removal, detention facilities, and 287(g) agreements, and that constitutional challenges to expanded interior expedited removal remained unresolved.
- Council on Foreign Relations, “How Trump Is Reshaping Immigration Enforcement”Supports claims about the administration’s enforcement expansion, more than 1,300 287(g) agreements by January 2026, major new enforcement funding, expanded expedited removal, and court challenges to third-country removals.
- Brennan Center for Justice, “How ICE’s Budget Boom Is Changing Immigration Detention”Supports claims about the scale of ICE detention funding, detention-capacity expansion, rising detainee numbers, and concerns about weakened oversight access.
What would change Advocate A · OpenAI GPT-5.5’s mind
I would change my mind if audited case-level data showed that expanded detention and removal capacity overwhelmingly executed valid final orders, produced low documented error rates, preserved counsel and court access, and remained subject to timely independent inspection and congressional access.
- The strongest pro-A claim is that lawful immigration tools can still create an accountability deficit when detention, expedited processing, and delegated enforcement expand faster than oversight can respond.
- 8 U.S.C. § 1225(b)(1) and 8 U.S.C. § 1357(g) authorize expedited removal and 287(g) agreements, so A must prove unlawful operation rather than merely statutory scale.
- ICE reported more than 7.6 million people on its non-detained docket at the end of FY2024 while removals totaled 271,484, indicating a large enforcement-capacity gap.
- Officially reported data stated that 1,445,549 non-detained people had final removal orders as of November 24, 2024.
- Documented detention deficiencies support stronger oversight remedies, but they do not by themselves prove that expanded enforcement capacity is extra-legal or beyond constraint.
Read the full argument
A’s strongest argument is that the word “shadow” need not mean secret or facially unlawful; it can mean an enforcement apparatus whose practical operation becomes difficult to inspect, contest, or correct before detention or removal occurs. On that view, the July 2025 funding surge, rapid detention expansion, wider 287(g) deputization, and interior expedited-removal policy create a real accountability risk, especially where reports describe serious facility deficiencies or diminished oversight access.
That is the best version of A’s case, and I accept its important caution: enforcement capacity can be abused, and lawful tools can produce unlawful applications. The specific reason B remains stronger is that A has not shown the decisive empirical step: that the expanded system is actually operating outside statutory constraints or that courts, inspectors, Congress, and counsel have been nullified rather than strained. A has shown scale, speed, and risk; B’s position requires a narrower but better-supported conclusion: these are congressionally authorized enforcement tools being expanded to address a demonstrable enforcement deficit.
The statutory baseline matters. Section 1225(b)(1) authorizes expedited removal for specified inadmissible noncitizens and includes asylum-related credible-fear screening and limited review; Section 1357(g) authorizes state-local immigration enforcement agreements under written agreements, training, federal supervision, and consistency with state and local law. These provisions may be too harsh as a policy matter, but that is a normative objection to Congress’s design, not proof of exceptional or shadow governance.
The administrative baseline also matters. ICE reported that its non-detained docket grew to more than 7.6 million by the end of FY2024 while removals totaled 271,484; a Federal Register public-inspection document reported 1,445,549 non-detained people with final removal orders as of November 24, 2024. A system that issues final orders but lacks capacity to execute them predictably undermines legality in the opposite direction: adjudication becomes symbolic. Detention beds, transport, hiring, and local cooperation are therefore not inherently suspicious; they are the infrastructure needed to make final legal judgments enforceable.
A’s oversight evidence deserves weight, but it cuts less decisively than A claims. AP’s reporting on Camp East Montana described 49 deficiencies, but the fact pattern also shows an ICE Office of Detention Oversight inspection, a posted report, press scrutiny, congressional criticism, and contractor replacement. That is not enough to declare the system healthy; it is enough to show that oversight mechanisms still generate actionable information. The better inference is not “shadow infrastructure,” but “lawful enforcement expansion requiring aggressive inspection, litigation, and congressional monitoring.” B is stronger because A’s argument proves a need for safeguards, not that enforcement capacity itself has become illegitimate.
Evidence cited by Advocate B · OpenAI GPT-5.5 (5) — as presented, not independently endorsed
- Legal Information Institute, 8 U.S.C. § 1225Supports that expedited removal is codified by Congress and includes specified procedures such as credible-fear screening and limited review.
- Legal Information Institute, 8 U.S.C. § 1357Supports that Congress authorized 287(g) agreements for trained state and local officers under written agreements and federal-law constraints.
- ICE FY2024 Annual Report releaseSupports that ICE removed 271,484 people in FY2024 and that the non-detained docket grew from more than 6.1 million to more than 7.6 million.
- Federal Register public inspection document, 2025-05120Supports the figure that 1,445,549 people on ICE’s non-detained docket had final orders of removal as of November 24, 2024.
- AP News, inspection of Camp East Montana detention facilitySupports that a congressionally mandated ICE Office of Detention Oversight inspection found 49 deficiencies at Camp East Montana and that the findings became public.
What would change Advocate B · OpenAI GPT-5.5’s mind
I would change my mind if audited case-level data showed that expanded removals systematically bypassed statutory eligibility limits, produced high wrongful-removal rates, or prevented timely court, counsel, congressional, and independent-inspector access in a large share of cases.
The Arbiter weighed this debate, verified the evidence, and took a position.
Read the article →