Government Acquiescence to Prison Gang Torture Under CAT: Commentary on B.G.S. v. Bondi (2d Cir. 2025)

I. Introduction

The Second Circuit’s decision in B.G.S. v. Bondi, No. 23‑6862 (2d Cir. Nov. 24, 2025), is a significant development in U.S. law under the Convention Against Torture (“CAT”). It clarifies how immigration adjudicators must analyze the risk of torture in foreign prisons when the feared torturers are private actors (such as gangs) and the state is alleged to be unable or unwilling to protect the applicant.

The petitioner, “B.G.S.,” is a Guatemalan national and former member of the Mara Salvatrucha gang (“MS‑13”). He sought deferral of removal under CAT, arguing that if returned to Guatemala he would likely be:

• Arrested on an outstanding murder warrant and imprisoned;
• Identified in prison as a former MS‑13 member who tried to leave and cover an MS‑13 tattoo;
• Brutally attacked or killed by MS‑13 or rival gangs in prison; and
• Left unprotected by Guatemalan officials who either collude with gangs or are effectively unable to prevent gang torture in prisons.

An Immigration Judge (“IJ”) and the Board of Immigration Appeals (“BIA”) (collectively, “the Agency”) denied CAT protection, concluding that while Guatemalan prisons are dangerous and violent, the conditions resulted from negligence or lack of resources, not from “state action” or government “acquiescence” as required by CAT. The Second Circuit majority (Judge Robinson, joined by Judge Kearse) granted the petition, vacated the BIA’s decision, and remanded. Judge Sullivan dissented.

The central legal issue is the proper application of the government acquiescence standard under CAT when the feared torture is by prison gangs, not by state officers themselves. The majority holds that the Agency misapplied the law by collapsing the question of unsafe prison conditions into the distinct inquiry whether officials would knowingly allow third‑party torture. The dissent believes the Agency did address that question and that its fact findings deserve deference.

II. Summary of the Opinion

A. Disposition

The Second Circuit:

• Granted the petition for review;
• Vacated the BIA’s denial of CAT deferral;
• Vacated the pending stay of removal (procedural point); and
• Remanded to the Agency for further proceedings consistent with the court’s opinion.

Importantly, the court did not grant CAT protection outright. Instead, it found legal error in the Agency’s analysis, requiring a new decision under the correct standard.

B. Core Holdings

The key holdings can be distilled as follows:

1. Distinct inquiries: prison conditions vs. acquiescence to third‑party torture.
   It is legally erroneous to limit the CAT analysis to whether a prison system’s poor conditions or generalized violence are themselves the result of intentional state torture. Where a petitioner specifically fears torture by private actors in prison (here, gangs), the Agency must separately evaluate:
   • Whether it is more likely than not that the petitioner will be tortured by those private actors; and
   • Whether government officials will, with knowledge or willful blindness, breach their duty to intervene—i.e., “acquiesce.”
2. Failure to apply the acquiescence standard to the prison‑gang scenario.
   The record showed that:
   • B.G.S. is likely to be arrested upon return (due to an outstanding warrant);
   • Guatemalan prisons are dominated by gangs, with extreme gang‑on‑gang brutality; and
   • His tattoos and history make him a distinctive target for gang violence.
   Yet the IJ and BIA treated the prison risk primarily as a problem of overcrowding and under‑resourcing. The majority held that the Agency failed to analyze whether prison officials would knowingly allow or be willfully blind to gang torture of B.G.S., contrary to governing precedent.
3. Guidance on the acquiescence analysis.
   On remand, the Agency must specifically consider:
   • Whether Guatemalan officials would know or be willfully blind to the risk of gang torture of B.G.S. in prison;
   • Whether they have a legal duty to intervene to protect him;
   • Whether, in practice, they are able and likely to prevent the torture; and
   • How to evaluate government efforts to combat gangs that coexist with corruption and partial complicity.
   The court reiterates that efforts by some officials to combat gangs do not automatically negate government acquiescence where others are complicit and the state is ineffective overall.
4. Credibility vs. persuasiveness.
   The panel flags that the IJ labeled witnesses (including B.G.S.) “credible but not persuasive.” Under Garland v. Ming Dai and Hong Fei Gao, credibility and persuasiveness are distinct concepts with procedural consequences. The court suggests that the IJ’s “persuasiveness” findings may in substance be adverse credibility determinations, and instructs the Agency to clarify this on remand.

C. The Dissent

Judge Sullivan dissented, concluding that:

• The IJ and BIA did in fact analyze whether Guatemala would acquiesce in prison‑gang torture, expressly finding that any harm to B.G.S. in prison would result from negligence and lack of resources, not willful blindness;
• That factual finding is supported by substantial evidence (including State Department reports and expert testimony about “iron fist” anti‑gang policies); and
• The court is bound to defer to the Agency’s factual findings unless no reasonable adjudicator could agree with them, which was not the case here.

Thus, in the dissent’s view, there was no legal error in the Agency’s application of the acquiescence standard, and the petition should have been denied.

III. Factual and Procedural Background

A. Petitioner’s Story

Key factual elements of the case include:

• Early recruitment into MS‑13.
  B.G.S. joined MS‑13 in Guatemala at around eight years old. He was tattooed with the emblem “La mano del Hueso” (“The hand of the Bone”) on his back, a symbol identifying him as an MS‑13 member.
• Leaving the gang and retribution.
  In his early 20s, he attempted to leave MS‑13 for the sake of his children. After informing local leader “Black Demon,” he was subjected to a ritual beating (13 members for 13 seconds) and then given a chance to reconsider. When he persisted in leaving, Black Demon put a “greenlight” on him—an order that other members could harm or kill him.
• Subsequent attacks in Guatemala.
  After he moved towns:
  • Three MS‑13 members attacked him with a machete, injuring him severely; a neighbor’s gunfire in the air disrupted the attack.
  • Later, Black Demon shot at him, and he sustained bullet wounds to his waist and abdomen.
  • He tried to cover the MS‑13 tattoo with an owl tattoo, but the original remained visible.
• Conflict with a powerful police officer.
  While attempting to isolate himself, he began a relationship with a woman who, unbeknownst to him at first, was also involved with A.S., a powerful local police officer who allegedly collaborated with MS‑13. A.S. twice shot at him (once injuring him under the arm) and reportedly drove around with uniformed officers in official vehicles searching for him.
• Flight through Mexico and into the United States.
  In 2020, he fled to Mexico, where two cartels threatened him and tried to recruit him. He then entered the United States without inspection around July 2021. In October 2021, he pled guilty to third‑degree assault in New York and served seven months in jail.
• Removal proceedings and CAT claim.
  Upon release, DHS charged him as removable. He applied for deferral of removal under CAT (having withdrawn asylum and statutory withholding claims), arguing that he faces torture in Guatemala at the hands of:
  • MS‑13, for leaving the gang and covering his tattoo;
  • A.S., the police officer whose girlfriend he had been involved with; and
  • Cartels that had threatened him in Mexico, now operating in Guatemala.
  Crucially, the government also introduced an outstanding Guatemalan arrest warrant for him on suspicion of murder, with sparse information about the underlying allegations.

B. Evidence of Country Conditions

Country‑conditions evidence, particularly from expert Jonathan Rosen and various reports, established:

• MS‑13’s treatment of deserters.
  MS‑13 views attempted exit and covering gang tattoos as severe disrespect. Former members who “defy their authority” are frequently harmed or killed. Tattoos “brand” individuals for life and make them easily visible targets.
• Gang‑dominated prisons and extreme violence.
  Multiple sources (USAID, State Department, Freedom House, academic studies) described Guatemalan prisons as:
  • Grossly overcrowded, harsh, and life‑threatening;
  • Effectively controlled in many respects by transnational gangs and drug trafficking groups;
  • Sites of extreme gang‑on‑gang violence, including torture, dismemberment, and high‑profile massacres (e.g., decapitation and burning of inmates).
• Government corruption and collusion.
  Reports documented:
  • Widespread penetration of criminal elements into security forces, judicial institutions, and prisons;
  • Instances where supposed crackdowns on gangs in prison were coordinated with gang leaders in advance;
  • A public perception that security forces are not merely ineffective but “downright malevolent.”
• Impact of the arrest warrant and tattoos.
  Rosen opined that:
  • Upon deportation, Guatemalan authorities would receive information about the outstanding warrant;
  • He would likely be classified as a gang member and targeted by police; and
  • His risk of torture or death upon return is “extremely likely.”

C. Agency Decisions

1. The IJ’s ruling

The IJ:

• Held a competency hearing and found B.G.S. competent, with safeguards (e.g., accepting a detailed affidavit instead of live direct examination, using simple questions, allowing breaks) due to trauma‑related mental health concerns.
• Found the testimony of:
  • B.G.S. credible but not persuasive, citing:
    • An inconsistency between his written statement (“shot at” by Black Demon) and live testimony (“shot” and actually hit); and
    • The IJ’s view that his explanation of how MS‑13 would find him upon return was speculative.
  • His mother credible but not persuasive, because she claimed not to know about his gang membership, had not personally witnessed threats, and was unsure if she wrote her own statement.
  • His brother credible but not persuasive, because his oral testimony that MS‑13 came to the family home conflicted with a written statement that did not identify the visitors as MS‑13.
  • Expert Rosen credible but not persuasive, deeming his testimony too generalized and insufficiently tailored to the specific facts of the case.
• Acknowledged medical evidence of scars consistent with bullet and machete wounds, but found it insufficient to connect those injuries specifically to MS‑13 retaliation for leaving the gang.
• Concluded that while MS‑13 and ex‑member violence is broadly documented, the record was too speculative to show it is more likely than not that MS‑13 or any gang would harm B.G.S. if he returned.
• Regarding prison:
  • Recognized that the outstanding warrant increased the likelihood of arrest and imprisonment;
  • Accepted evidence that Guatemalan prisons are harsh and life‑threatening; but
  • Found it “not clear from this record that the Guatemalan government is willfully blind to this harm rather than lacking resources,” and thus found no government acquiescence.
• Denied CAT protection after considering all evidence “in the aggregate.”

2. The BIA’s decision

On appeal, the BIA:

• Affirmed the IJ “for the reasons stated,” closely tracking the IJ’s reasoning;
• Expanded slightly on the prison analysis:
  • Recognized that prisons suffer from overcrowding, safety and control problems, and gang violence;
  • Characterized this as resulting from “negligence” and “lack of resources,” not “state action” sufficient to constitute “torture” under the regulations; and
  • Concluded that this also prevented a finding of government “acquiescence.”

IV. Legal Framework

A. The CAT Standard

The opinion reiterates the established two‑step framework for CAT claims, drawing on Chen v. Garland, 75 F.4th 109 (2d Cir. 2023), and Garcia‑Aranda v. Garland, 53 F.4th 752 (2d Cir. 2022):

1. Likelihood of torture (“more likely than not” test).
   The applicant must show it is “more likely than not” (i.e., more than a 50% chance) that he will be subjected to torture if removed. Under 8 C.F.R. § 1208.18(a)(1), “torture” means:

   any act by which severe pain or suffering, whether physical or mental, is intentionally inflicted on a person for such purposes as ... intimidating or coercing him or her or a third person ...

   The Agency must consider “[a]ll evidence relevant to the possibility of future torture,” including:
   • Evidence of past torture;
   • Ability to relocate internally;
   • Evidence of “gross, flagrant or mass violations of human rights”; and
   • Other relevant country‑conditions information. 8 C.F.R. § 1208.16(c)(3).
2. State action / government acquiescence.
   The applicant must then show that the torture would occur:

   at the instigation of or with the consent or acquiescence of a public official or other person acting in an official capacity.

   See 8 C.F.R. § 1208.18(a)(1). “Acquiescence” is defined in § 1208.18(a)(7) and interpreted by the Second Circuit (e.g., Scarlett v. Barr, 957 F.3d 316 (2d Cir. 2020); De La Rosa v. Holder, 598 F.3d 103 (2d Cir. 2010)) to mean:

   government officials know of or remain willfully blind to an act of torture and thereafter breach their legal responsibility to prevent it.

If both prongs are met, CAT relief is mandatory, not discretionary. See 8 C.F.R. § 1208.16(c)(4).

B. Deferral vs. Withholding of Removal Under CAT

CAT relief comes in two forms (8 C.F.R. § 1208.16(c)(4)):

• Withholding of removal: more durable protection; barred for individuals convicted of particularly serious crimes, serious non‑political crimes abroad, or those deemed national‑security risks. See 8 U.S.C. § 1231(b)(3)(B); 8 C.F.R. § 1208.16(d)(2).
• Deferral of removal: a “less permanent” and more easily terminable form of protection (64 Fed. Reg. 8480, 8481 (1999)), available even to those barred from withholding, including serious criminals. See 8 C.F.R. § 1208.17(a).

The standard of proof is the same for both; only the consequences differ. Here, because of his criminal conviction and the outstanding Guatemalan murder warrant, B.G.S. sought only deferral.

C. Standards and Scope of Judicial Review

The panel reviews:

• Factual findings for “substantial evidence” – they are conclusive unless “any reasonable adjudicator would be compelled to conclude to the contrary.” 8 U.S.C. § 1252(b)(4)(B); applied in Singh v. Bondi, 139 F.4th 189 (2d Cir. 2025);
• Questions of law and applications of law to fact de novo (no deference). See Hong Fei Gao v. Sessions, 891 F.3d 67, 76 (2d Cir. 2018);
• Both the IJ and BIA decisions “for the sake of completeness” where the BIA “closely tracks” the IJ’s reasoning, per Wangchuck v. DHS, 448 F.3d 524, 528 (2d Cir. 2006).

This case turns largely on whether the Agency misapplied the legal standard for acquiescence—a question the Second Circuit reviews de novo.

V. The Court’s Legal Analysis

A. The Core Error: Misapplying the Prison‑Gang Torture Framework

The majority agrees that the IJ correctly cited the principle from Pierre v. Gonzales, 502 F.3d 109 (2d Cir. 2007):

The failure to maintain standards of diet, hygiene, and living space in prison does not constitute torture under the CAT unless the deficits are sufficiently extreme and are inflicted ... intentionally rather than as a result of poverty, neglect, or incompetence.

But the majority insists this is only part of the relevant law. Pierre deals with when general prison conditions themselves amount to torture. B.G.S. presented a different, more specific theory:

• He is likely to be imprisoned due to a warrant;
• Inside prison, he faces targeted torture by private actors—MS‑13 or rival gangs—because of his status as an ex‑member and his distinctive tattoos;
• Prison officials, embedded in a system rife with gang control and corruption, will know of (or be willfully blind to) this risk and will not intervene.

The majority finds that the IJ and BIA never squarely applied the acquiescence standard to this prison‑gang scenario. Instead, they collapsed it into a general inquiry about overcrowding and violence attributable to lack of resources and negligence.

The court emphasizes that an acquiescence analysis requires distinct questions:

• Will the private actors (here, gangs) have the specific intent to inflict severe pain or suffering?
• Will government officials know of or be willfully blind to that likely torture?
• Do those officials have a legal duty to intervene and prevent the torture, and will they, in practice, breach that duty?

The majority holds that the Agency’s reasoning did not meaningfully engage those questions, particularly in light of substantial evidence of:

• Gang dominance within prisons;
• Documented extreme gang‑on‑gang brutality;
• Systemic corruption, including collusion of some officials with gangs; and
• The specific way that tattoos and attempted exit from MS‑13 mark B.G.S. as a target.

B. Use of Prior Precedents

1. Scarlett v. Barr (Jamaican police officer threatened by gangs)

In Scarlett, a former Jamaican police officer sought CAT protection after receiving death threats from gangs. The Second Circuit remanded because the Agency failed to:

• Assess the Jamaican authorities’ legal responsibility to protect a threatened officer; and
• Explain how their claimed inability to protect him related to the acquiescence standard.

B.G.S. draws a direct analogy: just as Jamaican authorities had a duty (and alleged inability) to protect Scarlett from gang violence, Guatemalan officials have duties to protect inmates from gang torture. The Agency’s failure to analyze those duties and their breach in the prison context echoes the error in Scarlett.

2. Garcia‑Aranda v. Garland (gang violence with police collusion)

In Garcia‑Aranda, the petitioner feared gang violence in Honduras with police participation or acquiescence. The Second Circuit remanded because the Agency did not address whether:

1. The gang would intentionally inflict severe pain (meeting CAT’s definition of torture); and
2. Local police would participate in or acquiesce in that torture.

Likewise, in B.G.S., the court notes that the Agency never clearly answered whether:

• Prison gangs are more likely than not to subject him to torture as defined in § 1208.18(a); and
• Prison officials will either participate in or acquiesce to that torture.

3. Pierre v. Gonzales (prison conditions vs. torture)

Pierre is a key foil. There, the court rejected a CAT claim based solely on harsh Haitian prison conditions absent targeted torture or intentional infliction of pain by officials.

The B.G.S. majority emphasizes that Pierre doesn’t foreclose CAT claims involving prison; it simply distinguishes:

• Generalized hardships due to poverty/neglect (usually not torture); from
• Specific acts of severe violence by private or state actors, with government knowledge and failure to intervene (potential torture under CAT).

Thus, relying on Pierre to treat all prison‑based claims as “just bad conditions” is a misapplication when the core allegation is third‑party torture with official acquiescence.

4. Khouzam, De La Rosa, and “partial” government efforts

The opinion cites Khouzam v. Ashcroft, 361 F.3d 161 (2d Cir. 2004), and De La Rosa v. Holder, 598 F.3d 103 (2d Cir. 2010), to reiterate that:

• Torture “requires only that government officials know of or remain willfully blind to an act and thereafter breach their legal responsibility to prevent it.” (Khouzam);
• The presence of some officials trying to fight gangs does not automatically preclude a finding of acquiescence, especially where:

  a government contains officials that would be complicit in torture, and that government, on the whole, is admittedly incapable of actually preventing that torture.

On remand, the Agency is expressly instructed to consider whether Guatemala’s anti‑gang efforts “override both the complicity of other government actors and the general corruption and ineffectiveness” in preventing prison gang violence. This underscores a nuanced view: anti‑gang initiatives and acquiescence can coexist.

5. Manning v. Barr (ignoring key evidence of likely death)

The court analogizes to Manning, 954 F.3d 477 (2d Cir. 2020), where the IJ and BIA overlooked critical evidence that the petitioner was likely to be killed if removed. The panel there held that ignoring such evidence required remand.

In B.G.S., there was substantial, specific evidence that:

• He would be imprisoned;
• He is “branded” by a partially visible MS‑13 tattoo and known exit from the gang;
• Gang‑dominated prisons are central arenas for exceptionally brutal gang‑on‑gang violence; and
• Former MS‑13 members and those who cover tattoos face extreme risk, particularly in prison.

The majority holds that the Agency’s failure to meaningfully confront this evidence—especially regarding how it affects the acquiescence analysis—is analogous to Manning.

6. Ming Dai and Hong Fei Gao: credibility vs. persuasiveness

In a notable footnote, the court highlights Garland v. Ming Dai, 593 U.S. 357 (2021), and Hong Fei Gao, 891 F.3d 67 (2d Cir. 2018), which distinguish:

• Credibility – whether testimony is believed, based on demeanor, consistency, plausibility, etc.;
• Persuasiveness – whether credible testimony, together with the rest of the evidence, is sufficient to meet the burden of proof.

An IJ’s adverse credibility finding requires “specific, cogent reasons” and triggers a presumption of credibility on appeal if not made. By contrast, a mere conclusion that testimony is “not persuasive” is different in law and effect.

Here, the IJ repeatedly stated that witnesses were “credible but not persuasive,” yet relied on inconsistencies (e.g., “shot” vs. “shot at”) in a manner that looks very much like classic credibility reasoning. The panel hints that, in substance, the IJ may have made adverse credibility findings without labeling them as such. On remand, the Agency is asked to clarify this, which may affect how the BIA and courts treat those findings.

C. Application to the Facts

The majority emphasizes that, even setting aside disputed aspects of personal testimony, there is independent, largely uncontested evidence showing:

• An outstanding warrant makes incarceration upon return likely;
• Guatemalan prisons are gang‑controlled environments with documented, extreme brutality;
• MS‑13 punishes deserters, particularly those who attempt to conceal gang tattoos;
• His tattoos and well‑known exit from MS‑13 make him easily identifiable as a target;
• There is systemic corruption and gang infiltration of law enforcement and prison systems.

Against this backdrop, the court concludes the Agency needed to address:

• Whether, in this environment, it is more likely than not that gang members will subject him to torture in prison; and
• Whether, given the evidence of gang dominance and corruption, Guatemalan officials can and will fulfill their duty to protect him, or instead will knowingly allow or be willfully blind to his torture.

Because the IJ and BIA addressed only the question of whether poor prison conditions themselves constitute state torture, and characterized all risk as negligence/lack of resources, the panel finds a failure of “reasoned consideration” on the key legal standard of acquiescence. This legal error necessitates remand.

VI. The Dissent’s Perspective

A. Core Argument

Judge Sullivan’s dissent contests the majority’s portrayal of the Agency’s analysis. In his view:

• The IJ expressly stated that “torturous conduct can either be committed directly by public officials or by private individuals [to whose actions] the authorities acquiesce” and quoted the correct definition of acquiescence, including awareness/willful blindness and breach of duty.
• The IJ specifically applied that standard to the prison context, concluding that violence in Guatemalan prisons “is not clear[ly]” the result of willful blindness rather than lack of resources, and that “the record does not demonstrate that the Guatemalan authorities would acquiesce to any potential torture.”
• The BIA then echoed and endorsed that finding, explicitly noting that “instances of violence . . . within Guatemalan prisons are the result of negligence,” not willfulness.

Thus, the dissent contends the Agency did analyze whether Guatemala would acquiesce to prison‑gang torture, and simply found that standard not met.

B. Deference and Substantial Evidence

The dissent underscores the strict standard for overturning factual findings: under 8 U.S.C. § 1252(b)(4)(B), those findings are conclusive unless any reasonable adjudicator would be compelled to find otherwise.

Judge Sullivan argues that the record supports the Agency’s view that:

• Guatemalan prison violence is attributable, at least in substantial measure, to negligence, lack of resources, and capacity constraints; and
• The government has pursued “iron fist” policies and tactics to combat gangs, even if often ineffective.

Because mistake, negligence, or even reckless disregard is not enough for “acquiescence” under Quintanilla‑Mejia v. Garland, 3 F.4th 569, 594 (2d Cir. 2021), the dissent believes the Agency’s no‑acquiescence finding is supported by substantial evidence and should be affirmed.

C. Analytical Divide

The disagreement between majority and dissent is thus less about the legal definition of acquiescence and more about:

• How to characterize the risk in Guatemalan prisons: pervasive negligence vs. systemic collusion and willful blindness;
• Whether the Agency’s reasoning shows it genuinely grappled with the prison‑gang‑torture theory, or merely re‑labeled it as generalized unsafe conditions; and
• How aggressive the court should be in demanding detailed, explicit analysis from the IJ and BIA in CAT cases.

VII. Impact and Broader Significance

A. Clarifying CAT Acquiescence in Prison‑Gang Contexts

B.G.S. v. Bondi is most important for its clarification of the government acquiescence analysis when torture is likely to occur in prison at the hands of gangs. It reinforces and extends prior cases like Scarlett and Garcia‑Aranda by:

• Requiring immigration adjudicators to distinguish clearly between:
  • General prison conditions (overcrowding, poor hygiene, etc.); and
  • Targeted acts of severe violence by private actors (gangs) against a specific individual.
• Insisting that the latter scenario be analyzed under the full acquiescence standard, even if the former is attributable to poverty or resource constraints.
• Emphasizing that evidence of:
  • Gang dominance in prisons;
  • Documented patterns of brutal gang‑on‑gang torture; and
  • Official corruption or collusion;
  can and should be central to the acquiescence analysis.

This will likely make it harder for the Agency, in future CAT cases, to dismiss prison‑based torture claims simply by characterizing all prison violence as the product of “negligence” or “lack of resources” without deeper engagement.

B. Implications for Former Gang Members and Tattooed Individuals

The case is particularly significant for:

• Former gang members who:
  • Have visible gang‑related tattoos;
  • Attempted to leave gangs that punish desertion harshly; and
  • Face deportation to countries where prisons are gang‑dominated.

The opinion recognizes, through the country‑conditions record, that:

• Tattoos both identify someone as linked to a gang and mark them as a traitor if altered or covered;
• Prisons may be more dangerous than the outside world for such individuals; and
• Gang targets may be singled out and subjected to particularly brutal forms of violence tantamount to torture.

Future petitioners in similar positions can cite B.G.S. to argue that the Agency must:

• Explicitly analyze the interplay between their tattoos, their status as defectors, and gang‑dominated prison environments; and
• Undertake a focused acquiescence inquiry rather than a generic prison‑conditions analysis.

C. Procedural and Evidentiary Practice Before IJs and the BIA

Practically, the decision sends several signals to immigration judges, the BIA, and counsel:

• Reasoned consideration of key evidence.
  IJs and the BIA must show they have “grappled” with:
  • Expert reports on prison gang structures;
  • Detailed country‑conditions evidence about corruption and collusion; and
  • Specific risk factors (tattoos, exit from gangs, outstanding warrants) that make a petitioner an identifiable target.
  A perfunctory or overly generalized reference to “overcrowding” or “lack of resources” may not suffice.
• Nuanced assessment of acquiescence.
  Decision‑makers must:
  • Identify the relevant public officials (e.g., prison wardens, guards);
  • Consider their knowledge or willful blindness to the torture risk; and
  • Assess the state’s capacity and actual practice of intervention.
• Careful use of “credibility” vs. “persuasiveness.”
  Post‑Ming Dai and Hong Fei Gao, IJs must:
  • Be explicit about whether they are making an adverse credibility finding or simply finding the evidence insufficiently persuasive;
  • Offer “specific, cogent reasons” if they doubt credibility;
  • Recognize that if no adverse credibility finding is made, the BIA must treat testimony as credible on appeal.
  B.G.S. underscores that using the label “not persuasive” to mask what are effectively credibility concerns may invite criticism and remand.

D. Potential for Doctrinal Divergence

While B.G.S. is firmly rooted in Second Circuit precedent, it also highlights ongoing tensions in CAT jurisprudence nationally:

• How to distinguish negligence/incompetence from willful blindness in deeply dysfunctional systems;
• How much weight to place on government “efforts” that are weak, fragmented, or offset by corruption;
• The extent to which generalized country conditions can, by themselves, prove that a specific petitioner is “more likely than not” to be tortured.

The majority’s insistence on a robust acquiescence analysis in the prison‑gang context could influence other circuits facing similar cases involving Central American gangs, African militias, or other non‑state prison actors.

VIII. Key Legal Concepts Simplified

A. What is “torture” under CAT?

“Torture” is not just any harm. Under CAT and its regulations, it means:

• Severe pain or suffering (physical or mental);
• Intentionally inflicted;
• For purposes such as punishment, coercion, intimidation, or discrimination; and
• With involvement or acquiescence of public officials acting in an official capacity.

Random acts of violence or very bad conditions—without this intentional, targeted quality and some level of state involvement—may not qualify as “torture” for CAT purposes.

B. “More likely than not”

The standard “more likely than not” means just what it sounds like: more than a 50% chance. Applicants do not need to prove certainty, but must show that torture is a probable outcome, not merely a possibility.

C. Government “acquiescence” and “willful blindness”

Even if the torturer is a private actor (e.g., a gang), CAT relief can be available if the government:

• Knows about the likely torture (or is willfully blind—deliberately avoids acknowledging it); and
• Has a legal duty to prevent it (e.g., to protect inmates) but fails to act.

“Willful blindness” is more than negligence. It implies that officials should know what is happening, see the warning signs, and choose not to look too closely or intervene.

D. Substantial evidence review

When a federal court reviews factual findings by the IJ and BIA, it applies a very deferential standard: it must uphold the findings unless “any reasonable adjudicator” would be forced to conclude otherwise. The court does not simply substitute its own view of the facts; it focuses on whether the Agency’s view is at least reasonable.

E. Credibility vs. persuasiveness

• Credible testimony – the judge believes the witness is telling the truth based on demeanor, consistency, and plausibility.
• Persuasive testimony – even if truthful, the testimony (alone or combined with other evidence) is strong enough to meet the legal burden of proof.

For example, a person may credibly testify about a single past beating, but that may not be persuasive enough to show they are “more likely than not” to be tortured in the future, especially if country conditions have changed.

Under current law, if an IJ does not explicitly find a witness not credible, the BIA must assume credibility on appeal. That makes it crucial for IJs to clearly distinguish and label their findings.

IX. Conclusion

B.G.S. v. Bondi is a major clarification of how U.S. courts expect immigration authorities to evaluate CAT claims involving prison‑based torture by gangs. The Second Circuit holds that:

• It is not enough to treat all prison violence and danger as a byproduct of poor resources or negligence; and
• Where the record shows likely incarceration, gang dominance, and a petitioner’s specific vulnerability (e.g., as a tattooed ex‑gang member), the Agency must undertake a focused analysis of whether officials will knowingly fail to protect him from torture.

By emphasizing the distinction between generalized prison conditions and targeted gang torture, and by insisting on a careful application of the acquiescence standard, the decision will shape how IJs and the BIA handle a broad class of CAT claims—particularly those involving former gang members and individuals marked by visible tattoos in gang‑dominated prison systems.

The dissent’s call for greater deference underscores the delicate balance between judicial oversight of legal standards and respect for the Agency’s fact‑finding role. Going forward, B.G.S. will likely serve as a leading authority in the Second Circuit on state acquiescence to private torture, especially within prisons, and will provide a blueprint for both advocates and adjudicators navigating the increasingly complex terrain of CAT protection.