No. 25-7052

Appeal from the United States District Court for the Eastern District of Oklahoma (D.C. No. 6:19-CV-00379-EFM)

Robert M. Blakemore (Daniel Smolen with him on the brief), Smolen & Roytman, Tulsa, Oklahoma, for Plaintiffs-Appellants.

Alison B. Levine (Wellon B. Poe with her on the brief), Collins Zorn & Wagner, PLLC, Oklahoma City, Oklahoma, for Defendant-Appellee.

Before HOLMES , Chief Judge, TYMKOVICH and MORITZ , Circuit Judges.

TYMKOVICH , Circuit Judge.

Page 2While incarcerated in an Oklahoma county jail, Shaun Smith and Savanaha Works alleged they were sexually abused by jailers.¹ After release, they filed a civil rights suit against the jailers for excessive force under 42 U.S.C. § 1983. They also alleged that the County Sheriff’s policies of understaffing the jail and inadequately supervising the jailers caused their abuse.

The district court granted summary judgment on the claim against the sheriff, finding no evidence that the jail was understaffed or that the sheriff was deliberately indifferent to the risk of unconstitutional sexual abuse. Exercising jurisdiction under 28 U.S.C. § 1291, we AFFIRM . A plaintiff asserting municipal liability based on a policy that does not facially violate federal law must prove that the municipality maintained the policy with deliberate indifference. This standard requires the municipality to have actual or constructive notice that its action or failure to act is substantially certain to result in an infringement on federal rights. The typical way to establish notice is through a pattern of similar constitutional or federal-statutory violations. Neither Smith nor Works has introduced evidence of a pattern of unconstitutional sexual abuse at the jail. As a result, we find the Sheriff’s Office was not deliberately indifferent in maintaining its staffing and supervision policies, and the municipal liability claims therefore fail.

Shaun Smith entered the Pushmataha County Jail in February 2016 for pretrial detention on a burglary charge. On March 3, the Pushmataha County Sheriff’s Office hired Tamara Nichols as a jailer and assigned her to the night shift. Nichols had worked at the jail twice before and had been dismissed for various reasons. The Sheriff’s Office first hired Nichols in 2012, fired her in 2013 for repeatedly missing work, rehired her in 2014, and fired her again after she tested positive for methamphetamine. Nichols also had several consensual sexual encounters with two previous inmates. The encounters typically occurred in the jail laundry room, which lacks security cameras because it is the room where inmates change from street clothes into jail clothes.

The Pushmataha County Sheriff at the time, Terry Duncan, was aware of Nichols’s previous sexual relationships. Though the jail had a sexual misconduct policy prohibiting any sexual contact between jailers and inmates, Duncan did not discipline Nichols. At some point in March or April 2016, Smith and Nichols engaged in a sexual relationship. Smith argues the encounter was nonconsensual. He says he has an IQ of 57, suffers from mental illness, and has abused drugs in the past. And he argues Nichols had been using him as a conduit to smuggle drugs into the prison and used that as leverage to manipulate him into nonconsensual sex.

Page 4A few weeks after the incident, another jailer saw Nichols and Smith rummaging through inmate property on the jail’s video surveillance system. The jail then asked the Oklahoma State Bureau of Investigation (OSBI) to investigate potential theft and mishandling of property. During her interview with an investigator, Nichols disclosed that she had sex with Smith in the laundry room. Nichols was arrested the same day, charged with second-degree rape, and terminated for violating the jail’s sexual misconduct policy.

Sheriff Duncan lost his bid for reelection in June 2016 and resigned from office in July. B.J. Hedgecock took over as sheriff in November and, aware of the Nichols–Smith incident, immediately worked to improve conditions at the jail. Sheriff Hedgecock fired several jailers, brought in new staff and officers, replaced the jail administrator, and asked the Oklahoma Jail Inspection Division to conduct an inspection and identify deficiencies he could address. He also revamped the sexual misconduct policy and enforced it. But one policy Hedgecock left in place from the previous administration was the practice of single staffing the jail during the overnight shift.

Savanaha Works was an inmate in the jail from July 2017 to November 2017. The Sheriff’s Office hired Timothy Byers as a jailer in October 2017 and assigned him to the overnight shift. After “lights out” on November 13, Byers told Works he needed her to get some jumpsuits and paper from the laundry room. Works’s role as Page 5a prison trustee afforded her more freedom of movement and allowed her to perform work around the prison. Byers followed Works to the laundry room and eventually told her to “drop them” and touch her toes. App. 441. He then initiated sexual contact for about five minutes.

The morning after the incident, another inmate reported that Byers removed Works from her cell and “forced her to have sex with him.” App. 318. Sheriff Hedgecock contacted OSBI and requested an investigation. Interviews with inmates produced conflicting stories about the event but confirmed a sexual encounter. A security camera outside the laundry room partially captured the incident and reveals Byers standing behind Works for about five minutes and likely engaging in some kind of sexual contact.

Works claimed that the encounter was coerced, she was too scared to call for help during the incident, and she did not report it out of fear. When investigators asked Byers whether the encounter was consensual, he invoked the Fifth Amendment and declined to answer. Hedgecock fired Byers, but because jail policy prohibits sex between inmates and jailers regardless of consent, the Sheriff’s Office made no determination on whether Works consented. Byers was later arrested and charged with second-degree rape by instrumentation, but the charge was dismissed after Works refused to cooperate with the prosecution.

Smith and Works jointly sued Nichols and Byers under § 1983 for allegedly subjecting them to excessive force and failing to protect them in violation of the Page 6Eighth and Fourteenth Amendments. They also brought a municipal liability claim by suing Sheriff Hedgecock in his official capacity, alleging he and his predecessor caused the sexual abuse through deliberate indifference by inadequately staffing the jail and failing to train and supervise the jailers.

Nichols moved to dismiss based on untimely service of process. The district court granted the motion after Smith and Works did not reply. Byers then moved for summary judgment based on qualified immunity, arguing there was no underlying constitutional violation because Works had consented to sex. The district court denied summary judgment because it determined Works’s consent was a contested fact issue and a prisoner’s right to be free from nonconsensual sexual contact was clearly established. Byers appealed, and we affirmed, holding Works had introduced sufficient evidence of nonconsent to defeat qualified immunity at summary judgment. Works v. Byers , 128 F.4th 1156, 1163 (10th Cir. 2025).

Sheriff Hedgecock also moved for summary judgment, arguing there were no grounds for municipal liability because there was no underlying constitutional violation. Alternatively, he asserted that no official policy caused the alleged constitutional violations and neither he nor Duncan had acted with deliberate indifference to an obvious risk of inmate sexual abuse. The district court recounted its finding that Works had created a material fact issue over her consent and then determined Smith had done the same regarding his consent. It therefore held, for summary judgment purposes, that both Smith and Works had been sexually abused in violation of their constitutional rights. Even so, the court ruled for Sheriff Page 7Hedgecock based on its findings that the prison was adequately staffed across both administrations, that both Nichols and Byers were trained on jail rules and state law governing sexual contact between inmates and jailers, and that neither Duncan nor Hedgecock had been deliberately indifferent to an obvious risk of unconstitutional sexual abuse of inmates.

On remand from Works’s appeal, Works and Byers settled and stipulated to Byers’s dismissal from the case. That rendered the grant of summary judgment for Sheriff Hedgecock a final order eligible for appellate review. Works and Smith timely appealed.

Smith and Works appeal the district court’s denial of their municipal liability claims based on inadequate jail staffing and failure to supervise.² We reject their arguments because, even assuming their constitutional rights were violated, they cannot prove that either Sheriff Duncan or Sheriff Hedgecock acted with deliberate indifference to an obvious risk that jailers would sexually abuse inmates.

“We review a district court’s grant of summary judgment de novo, using the same standard applied by the district court pursuant to Fed. R. Civ. P. 56(a).” Cillo v. City of Greenwood Village , 739 F.3d 451, 461 (10th Cir. 2013) (citation modified). At this stage, “[w]e must ‘view facts in the light most favorable’ to the non-moving Page 8part[y], . . . resolving all factual disputes and reasonable inferences in [his] favor.” Id. (quoting Tabor v. Hilti, Inc. , 703 F.3d 1206, 1215 (10th Cir. 2013)). “Summary judgment must be granted if ‘there is no genuine dispute as to any material fact’ and the moving party is ‘entitled to judgment as a matter of law.’” Id. (quoting Fed. R. Civ. P. 56(a)).

When a plaintiff sues a municipal employee in his official capacity under § 1983, the suit is an action “against the county or municipality.” Porro v. Barnes , 624 F.3d 1322, 1328 (10th Cir. 2010). While § 1983 allows claims against municipal entities for violating federal rights, liability is strictly limited to instances when the violation stems from “official municipal policy.” Monell v. Dep’t of Soc. Servs. of City of New York , 436 U.S. 658, 691 (1978). Thus, on top of an underlying constitutional or federal statutory violation, a municipal-liability plaintiff must prove three elements: (1) an official policy or custom, (2) that caused the underlying injury, and (3) culpable state of mind. Schneider v. City of Grand Junction Police Dep’t , 717 F.3d 760, 769 (10th Cir. 2013). When the plaintiff does not claim “the municipal action itself violated federal law, or directed or authorized the deprivation of federal rights,” the necessary state of mind is “deliberate indifference.” Bd. of Cnty. Comm’rs of Bryan Cnty. v. Brown , 520 U.S. 397, 406–07 (1997).

A municipality may be liable for an employee’s violation of a person’s federal rights when the claim is based on a formal policy or custom. Policies meeting this standard include, among others, “decisions of municipal employees with final Page 9policymaking authority” and “the deliberately indifferent failure to adequately train or supervise employees.” Pyle v. Woods , 874 F.3d 1257, 1266 (10th Cir. 2017) (citing Brammer-Hoelter v. Twin Peaks Charter Acad. , 602 F.3d 1175, 1189 (10th Cir. 2010)); see also Timothy M. Tymkovich, Municipal Liability: Tensions in the Tenth Circuit , 100 Denv. L. Rev. 439, 444–58 (2023) (identifying and categorizing types of municipal liability claims). “Under Oklahoma law, the sheriff is the final policymaker for a county jail.” Est. of Crowell ex rel. Boen v. Bd. of Cnty. Comm’rs of Cnty. of Cleveland , 237 P.3d 134, 142 (Okla. 2010).

“To establish the causation element, the challenged policy must be ‘closely related to the violation of the plaintiff’s federally protected right.’” Schneider , 717 F.3d at 770 (quoting Martin A. Schwartz, Section 1983 Litigation: Claims & Defenses § 7.12[B] (2013)). This means the plaintiff must show the policy 'was the ‘moving force’ behind the injury alleged.” Bd. of Cnty. Comm’rs of Bryan Cnty. , 520 U.S. at 404. One way to meet this burden is by demonstrating the policy facially violates federal law. See Barney v. Pulsipher , 143 F.3d 1299, 1307 (10th Cir. 1998). But when a policy is facially lawful, “the burden of establishing causation . . . is heavy.” Hinkle v. Beckham Cnty. Bd. of Cnty. Comm’rs , 962 F.3d 1204, 1241 (10th Cir. 2020) (citing Bd. of Cnty. Comm’rs of Bryan Cnty. , 520 U.S. at 405 (“Where a plaintiff claims the municipality has not directly inflicted an injury, but nonetheless caused an employee to do so, rigorous standards of culpability and causation must be applied to ensure that the municipality is not held liable solely for the actions of its employee.”)).

Page 10“[A] plaintiff seeking to establish municipal liability on the theory that a facially lawful municipal action has led an employee to violate a plaintiff’s rights must demonstrate that the municipal action was taken with ‘deliberate indifference’ as to its known or obvious consequences.” Bd. of Cnty. Comm’rs of Bryan Cnty. , 520 U.S. at 407 (quoting City of Canton v. Harris , 489 U.S. 378, 388 (1989)). This requires more than “[a] showing of simple or even heightened negligence.” Id. The standard “may be satisfied when the municipality has actual or constructive notice that its action or failure to act is substantially certain to result in a constitutional violation, and it consciously or deliberately chooses to disregard the risk of harm.” Hinkle , 962 F.3d at 1241 (quoting Waller v. City and County of Denver , 932 F.3d 1277, 1284 (10th Cir. 2019)). Such notice is typically established by proving municipal actors were aware of a pattern of constitutional violations but can also be proven “in a narrow range of circumstances where a violation of federal rights is a highly predictable or plainly obvious consequence of a municipality’s action or inaction.” Waller , 932 F.3d at 1284 (citation modified).

Works claims Byers violated her Eighth and Fourteenth Amendment rights by sexually abusing her and failing to protect her while she was imprisoned. A guard’s sexual abuse of an inmate violates the Constitution because sexual abuse “has no legitimate penological purpose, and is ‘simply not part of the penalty that criminal offenders pay for their offenses against society.’” Giron v. Corr. Corp. of Am. , 191Page 11 F.3d 1281, 1290 (10th Cir. 1999) (quoting Boddie v. Schnieder , 105 F.3d 857, 861 (2d Cir. 1997)).

We have recognized, however, “that consent is a defense to a constitutional claim for sexual abuse.” Brown v. Flowers , 974 F.3d 1178, 1186 (10th Cir. 2020) (citing Graham v. Sheriff of Logan Cnty. , 741 F.3d 1118, 1125–26 (10th Cir. 2013)). That is so because “not all misbehavior by public officials, even egregious misbehavior, violates the Constitution.” Graham , 741 F.3d at 1125. As we explained in Graham , the Supreme Court has cautioned us against constitutionalizing all tortious conduct by government actors. Id. (first citing DeShaney v. Winnebago Cnty. Dep’t of Soc. Servs. , 489 U.S. 189, 202 (1989) (“[T]he claim here is based on the Due Process Clause of the Fourteenth Amendment, which, as we have said many times, does not transform every tort committed by a state actor into a constitutional violation.”); and then citing Baker v. McCollan , 443 U.S. 137, 146 (1979) (“Section 1983 imposes liability for violations of rights protected by the Constitution, not for violations of duties of care arising out of tort law. Remedy for the latter type of injury must be sought in state court under traditional tort-law principles.”)). As a result, we found it “proper to treat sexual abuse of prisoners as a species of excessive-force claim, requiring at least some form of coercion (not necessarily physical) by the prisoner’s custodians.” Id. at 1126. Importantly, we have also emphasized that courts should be attentive to the high risk of nonphysical coercion based on the hierarchical relationship between guards and prisoners, and the inherently restrictive and disciplinary nature of jails. See id. (“We agree with the Page 12Ninth Circuit that ‘[t]he power dynamics between prisoners and guards make it difficult to discern consent from coercion.’” (quoting Wood v. Beauclair , 692 F.3d 1041, 1047 (9th Cir. 2012))); Brown , 974 F.3d at 1186–87 (considering “the inherently coercive nature of prisons” when analyzing whether a prisoner consented to sex with a jailer). Nonetheless, “in this circuit, the burden remains on the plaintiff—not the defendant—to establish that sexual conduct is nonconsensual.” Works , 128 F.4th at 1162.

Sheriff Hedgecock argues Works’s rights were not violated because she consented to sex with Byers. As discussed earlier, in a previous appeal we have already considered Works’s allegations and held “a reasonable jury could accept as true that Byers used nonconsensual force to invade Works’[s] bodily integrity.” Id. at 1165. We thereby affirmed the district court’s summary judgment determination that Byers violated Works’s constitutional rights. We see no reason to revisit that decision.³ As a result, Works has established a constitutional violation for purposes of her municipal liability claim.

Page 13Similarly, Sheriff Hedgecock contends Smith’s rights were not violated because he consented to sex with Nichols. The district court found the issue of Smith’s consent was a disputed fact and proceeded to the municipal liability analysis. Given that finding and the significant factual and legal overlap between Smith’s and Works’s municipal liability claims, we too assume Smith has established an underlying constitutional violation and move on to municipal liability.

Works argues for municipal liability based on Sheriff Hedgecock’s alleged failures to staff the prison adequately and to supervise Byers. Smith makes a similar argument based on Sheriff Duncan’s continuation of the allegedly insufficient staffing policy and inadequate supervision of Nichols. We have recognized both inadequate jail staffing and failure to supervise as official policies that can support municipal liability. See Lopez v. LeMaster , 172 F.3d 756, 763 (10th Cir. 1999) (discussing inadequate jail staffing and failure to supervise in the context of a municipal failure-to-protect claim); see also Prince v. Sheriff of Carter Cnty. , 28 F.4th 1033, 1049–50 (10th Cir. 2022) (denying summary judgment based on evidence that a jail’s deliberately indifferent failure to retain any medical staff contributed to a prisoner’s death). Importantly, while jail staffing can be so inadequate that it violates minimum constitutional standards, see Lopez , 172 F.3d at 763, neither Smith nor Page 14Works advanced that theory before the district court.⁴ Instead, they argued that the sheriffs’ one-jailer-per-shift policy fell short of Oklahoma state standards for staffing and allowed Nichols and Byers to perpetrate the alleged sexual abuse. See App. 723– 25.

But evidence of state law violations, even criminal misconduct, does not show a violation of federal law actionable under § 1983. See Graham , 741 F.3d at 1125– 26 (explaining that § 1983 protects against violations of constitutional rights, not common torts and criminal misconduct); Fee v. Herndon , 900 F.2d 804, 808 (5th Cir. 1990) (“[T]he Constitution is not a criminal or civil code to be invoked invariably for the crimes or torts of state educators who act in contravention of the very laws designed to thwart abusive disciplinarians.”). Since Smith and Works do not claim that the staffing policy violated federal law itself, they must prove the sheriffs maintained it with deliberate indifference. That same culpability standard applies to their failure-to-supervise claim. See Brammer-Hoelter , 602 F.3d at 1190.

Page 15Smith and Works point to no evidence that either Sheriff Duncan or Sheriff Hedgecock acted with deliberate indifference.⁵

Beginning with Sheriff Duncan, Smith argues that Duncan should have known that inadequate staffing and supervision “was substantially certain to result in a constitutional violation.” Opening Br. 28. It is true that “deliberate indifference may be found absent a pattern of unconstitutional behavior if a violation of federal rights is a ‘highly predictable’ or ‘plainly obvious’ consequence of a municipality’s action or inaction.” Barney , 143 F.3d at 1307–08 (first quoting Bd. of Cnty. Comm’rs of Bryan Cnty. , 520 U.S at 409; and then citing City of Canton , 489 U.S. at 390 & n.10). But the “range of circumstances” that can satisfy this standard is “narrow,” id. at 1308, and we have previously rejected the argument that prisoner sexual abuse is a plainly obvious consequence of having a single jailer on duty, id. at 1309 n.8 (citing Hovater v. Robinson , 1 F.3d 1063, 1066 (10th Cir. 1993)).

Smith next argues that if a pattern of prior incidents is necessary, Sheriff Duncan had notice of a substantial risk that Nichols would sexually abuse inmates based on her prior sexual misconduct. As we describe above, before the episode with Smith, the record shows that Nichols had maintained sexual relationships with two other inmates. Both relationships took place despite the jail’s misconduct policy prohibiting sex between jailers and inmates. Further, Oklahoma law criminalizes Page 16such conduct by a jailer as statutory rape. Okla. Stat. tit. 21, § 1111(7) (2015). Sheriff Duncan was aware of the relationships and even confronted one of the inmates, who simply laughed and walked away.⁶ According to Smith, the fact that Sheriff Duncan knew Nichols had repeatedly violated prison rules and state law governing sexual contact between inmates and jailers means he was aware of the risk of sexual abuse presented by inadequate staffing and supervision.

Smith’s argument fails because even deplorable prior violations of prison rules and state law do not establish a pattern of infringement against federal rights. See Bd. of Cnty. Comm’rs of Bryan Cnty. , 520 U.S. at 409 (explaining notice ordinarily requires “showing a pattern of constitutional violations”); cf. Schneider , 717 F.3d at 774 (holding police officers’ consensual sexual relationships with private citizens they met while on duty, though a violation of department policy, did “not amount to constitutional violations”). Since our precedent holds that uncoerced sex between a jailer and inmate is not a constitutional violation, see Brown , 974 F.3d at 1186, Smith must prove a pattern of nonconsensual sex at the jail to establish that Sheriff Duncan had notice that his policies would lead to sexual abuse. Smith does not make an argument that the relationship was nonconsensual.

Page 17As we noted above, the line between consent and coercion in this context is hard to draw, and our concerns about “the power dynamics between prisoners and guards” have led us to scrutinize claims by jailers that prisoners consented to sex. Graham , 741 F.3d at 1126 (citation modified); see also Brown , 974 F.3d at 1186. But we recognized in Graham that a prisoner's claim of sexual abuse certainly fails when there is no genuine dispute over consent. 741 F.3d at 1123. The evidence of Nichols’s conduct with one of the inmates satisfies that standard because the inmate testified that all the encounters were consensual. App. 877–78; see Brown , 974 F.3d at 1186–87 (discussing the role of testimony in establishing consent or nonconsent). Similarly, Nichols claims that her relationship with a second prior inmate was consensual, and Smith and Works have not identified any contradictory evidence. Thus, the only prisoner–jailer relationships that Duncan knew about were apparently consensual and therefore not constitutional violations.

Pointing to our decision in Tafoya v. Salazar , 516 F.3d 912 (10th Cir. 2008), Smith suggests that prison officials’ failure to enforce a no-contact policy can, by itself, prove deliberate indifference. But that misunderstands the case. In Tafoya , a sheriff implemented a no-contact policy between male officers and female inmates after a series of known sexual assaults at a prison under his control. Despite the new policy, a male guard sexually assaulted a female inmate three years later. When the inmate sued and alleged deliberate indifference, the sheriff pointed to the no-contact policy as evidence that he had worked to improve conditions and protect inmates from abuse and was thus not deliberately indifferent. We rejected the sheriff’s Page 18argument because there was testimony that he had not enforced the policy and knew that jailers routinely violated it. Instead, we drew an inference of deliberate indifference based on the sheriff’s notice of the prior assaults and his failure to enforce his own mitigation measures despite that notice. The pattern of prior constitutional violations was key to our decision because it showed the sheriff was aware of a specific risk of more violations and made only a faithless effort at improvement. But in Sheriff Duncan’s case, the requisite notice is absent. Without such notice, we cannot infer that his non-enforcement of the jail’s sexual misconduct policy was undertaken with deliberate indifference.

No doubt, Sheriff Duncan’s lax enforcement of rules and willful disregard for criminal sexual conduct inside the jail is inexcusable. But evidence that Nichols engaged in a consensual sexual relationship with a prisoner, while outrageous, does not indicate a high likelihood that Nichols would also force an inmate into sex against his will. Since Smith has not introduced evidence of any constitutional violation prior to his abuse, let alone a pattern of them, he cannot prove that Sheriff Duncan acted with deliberate indifference.

Thus, his municipal liability claim fails.

Works similarly argues that Sheriff Hedgecock was deliberately indifferent because he was aware of the prior sexual misconduct between Smith and Nichols and chose not to change the staffing policy. She also claims that Hedgecock knew that double-staffing with jailers of each sex would make sexual abuse much less likely. According to Works, the combination of these facts proves that Hedgecock’s Page 19persistence in single-staffing the jail was done with deliberate indifference to the risk of prisoner sexual abuse. We disagree.

Once again, sexual misconduct under prison rules and state law is not evidence of a constitutional violation. See Graham , 741 F.3d at 1125–26. But even assuming that Smith was sexually abused would not change the outcome. As we noted earlier, prisoner sexual abuse is not a plainly obvious consequence of a single-staffing policy. Barney , 143 F.3d at 1309 n.8 (citing Hovater , 1 F.3d at 1066). So we are outside the “narrow range of circumstances” where a single prior incident is adequate to give notice of a policy deficiency. Id. at 1307 (quoting Bd. of Cnty. Comm’rs of Bryan Cnty. , 520 U.S. at 409). Further, the Smith–Nichols incident did not establish a pattern of similar constitutional violations that would alert Sheriff Hedgecock to a problem with his staffing policy. See Waller , 932 F.3d at 1287 (“[O]ne prior incident, even if it was a constitutional violation sufficiently similar to put officials on notice of a problem, does not describe a pattern of violations.” (quoting Coffey v. McKinley Cnty. , 504 F. App’x 715, 719 (10th Cir. 2012))). Thus, Hedgecock’s knowledge of that incident did not make it plainly obvious that another, similar constitutional violation would result from continuing to single-staff the jail. Nor does his acknowledgment of the potential virtues of double-staffing tip the scales. At most, it shows Sheriff Hedgecock knew that increased staffing was one possible mitigation tool among many. And he acted to abate the risk of prisoner abuse, including through strict enforcement of the sexual misconduct policy. Whatever the theoretical benefits of double-staffing, absent some notice that single-staffing would Page 20lead to sexual abuse, Sheriff Hedgecock did not act with deliberate indifference by choosing to implement different mitigation measures.

Without notice that his staffing and supervisory policies carried a high probability of causing jailers to sexually abuse prisoners, Sheriff Hedgecock’s maintenance of those policies was not done with deliberate indifference. Works has not introduced evidence sufficient to find that Sheriff Hedgecock had the necessary notice. As a result, her municipal liability claim fails.

For the foregoing reasons, we affirm the district court’s judgment.

MORITZ , Circuit Judge, concurring.

I join the majority opinion because it faithfully applies our precedent in Graham v. Sheriff of Logan County , which implicitly adopted a weak presumption— but a presumption nevertheless—that inmates consent to sexual encounters with correctional officers absent “some form of coercion.” 741 F.3d 1118, 1126 (10th Cir. 2013); see also Works v. Byers , 128 F.4th 1156, 1162 (10th Cir. 2025) (explaining that the burden is on plaintiff, not defendant, “to establish that sexual conduct is nonconsensual”). But prison, as we have recognized, is an inherently coercive environment. Works , 128 F.4th at 1160; see also Brown v. Flowers , 974 F.3d 1178, 1187 (10th Cir. 2020). Presuming consent in such an environment legally sanctions institutional coercion and transforms a plain constitutional violation into a permissive interaction. Instead, we should adopt a straightforward presumption that inmates do not consent to sexual activity with prison staff, and that such sexual activity violates the Eighth Amendment, absent compelling evidence of consent put forth by the defendant.

At the same time, this appeal is not well-suited to reconsidering Graham : plaintiffs have not asked us to do so, nor have they advanced the arguments I present below. See Clark v. Sweeney , 607 U.S. 7, 9 (2025) (explaining that our adversarial system “follow[s] the principle of party presentation,” meaning the parties frame the issues and the court decides them (quoting United States v. Sineneng-Smith , 590 U.S. 371, 375 (2020))). Nevertheless, this appeal offers an opportunity to evaluate Page 22Graham both in the municipal-liability context and in light of evolving standards of decency and relevant case law. See, e.g. , Walton v. Nehls , 135 F.4th 1070, 1075 (7th Cir. 2025) (noting that federal law and “all 50 states have made a prison official’s sexual conduct with a prisoner a crime”). So I take this opportunity to write separately, recognizing that this will not be the last time we will be asked to review cases arising from sexual encounters between inmates and correctional officers.

As the majority explains, Graham held that consent is a defense to inmate claims of sexual assault asserted under the Eighth Amendment, rejecting the plaintiff’s argument that prisoners can’t legally consent to sex with their custodians.¹ 741 F.3d at 1124. To do so, Graham first surveyed the case law, noting that the Sixth and Eighth circuits permitted a consent defense, several district courts did not, and the Ninth Circuit followed a “middle ground” of presuming nonconsent. Id. at 1124– 25. Graham thus found “no consensus in the federal courts on whether, or to what extent, consent is a defense to an Eighth Amendment claim based on sexual contact with a prisoner.” Id. at 1125. So, falling back on warnings from the Supreme Court “against constitutionalizing . . . tortious conduct by government agents,” Graham Page 23held consent was a defense but declined to adopt a presumption of nonconsent.² Id. at 1125–26.

Viewing Graham through the lens of Smith’s municipal-liability claim reveals some of its limitations. As the majority accurately explains, Smith’s municipalliability claim requires a showing of deliberate indifference, a standard typically satisfied by showing that officials failed to act in the face of a pattern of constitutional violations. Here, Smith showed that Sheriff Duncan failed to act in the face of a pattern of violations of prison policy and state law—violations that bear an exceedingly close resemblance to the conduct underlying Smith’s claim. But “[s]ince our precedent holds that uncoerced sex between a jailer and inmate is not a constitutional violation,” that’s not enough. Maj. Op. 16. Instead, “Smith must prove a pattern of nonconsensual sex at the jail to establish that Sheriff Duncan had notice that his policies would lead to sexual abuse.” Id. I agree with the majority that Smith did not attempt to make such a showing here.

Page 24That leaves us in the untenable position of permitting the municipality to avoid § 1983 liability on Smith’s sexual-assault claim even though Sheriff Duncan concedes knowing about Nichols’s prior and repeated sexual activity with multiple inmates in violation of prison policy and state law. And it’s worth noting that Sheriff Duncan not only chose to overlook this repeated sexual activity, he directed the inmate— not his employee — to knock it off. To be sure, violations of policy and state law don’t inherently rise to the level of constitutional violations, and I respect the Supreme Court’s admonitions “against constitutionalizing all tortious conduct by government actors.” Id. at 11. In the abstract, the idea that consensual sexual encounters are tortious, but not cruel and unusual, is not necessarily objectionable. But given the inherently coercive nature of the relationship between correctional officers and inmates, cases falling into that category are likely to be exceedingly rare. Freely given, uncoerced consent in the prison environment is highly unlikely because “prisoners depend on prison officials ‘for nearly everything in their lives—their safety as well as their access to food, medical care, recreation, and even contact with family members.’” Walton , 135 F.4th at 1075 (quoting J.K.J. v. Polk County , 960 F.3d 367, 381 (7th Cir. 2020)).

Given these dynamics, the risk of misinterpreting a nonconsensual sexual encounter as consensual is far higher—and far more damaging to Eighth Amendment rights—than the alternative, misinterpreting a consensual encounter as nonconsensual. As Graham itself acknowledged, “‘the power dynamics between prisoners and guards make it difficult to discern consent from coercion.’” 741 F.3d atPage 25 1126 (quoting Wood , 692 F.3d at 1047). And one can imagine many reasons an inmate might testify that past sexual encounters were consensual, even though they were not. In fact, I would hesitate before relying wholesale on inmate testimony to the exclusion of other evidence. “[F]avors, privileges, or any type of exchange for sex” can “indicate coercion” as much as “explicit assertions or manifestations of non[]consent.” Wood , 692 F.3d at 1049; see also Brown , 974 F.3d at 1185 (explaining that our case law “do[es] not delineate between sexual abuse carried out through physical and nonphysical coercion').³

Even assuming consensual and nonconsensual sexual encounters are readily distinguishable in the prison context, permitting sexual encounters between staff and inmates creates a prison environment in which constitutional violations in the form of nonconsensual sex are much more likely. Our current case law nevertheless insulates a municipality from § 1983 liability when its officials fail to correct, and even tolerate, that environment. As Smith puts it, Nichols’s prior “sexual misconduct [wa]s highly probative” of the nature of her conduct toward Smith, and Sheriff Duncan’s nonplussed response to Nichols’s prior misconduct shows “an abject failure to supervise.” Aplt. Br. 25–26. Stated plainly, common sense suggests that Sheriff Duncan’s failure to investigate, discipline, or fire Nichols constituted deliberate indifference to sexual misconduct. Common sense would in turn suggest that Sheriff Page 26Duncan was deliberately indifferent to an increased risk of nonconsensual sexual activity in his jail, an Eighth Amendment violation. But the law does not permit municipal liability because all Sheriff Duncan had notice of was presumably consensual sex, which means he only had notice of a pattern of tortious (or criminal), rather than unconstitutional, behavior.

Adopting a presumption of nonconsent would go a long way toward rectifying this mismatch between common sense and legal doctrine. It would acknowledge the rarity of truly consensual sexual encounters in the prison setting. And it would address the difficulty of discerning consensual from nonconsensual sexual activity by placing the burden on prison staff and municipalities to prove consent, thereby creating an incentive to enforce zero-tolerance policies, in turn curtailing violations of prisoners’ Eighth Amendment rights.

In this case, starting with the presumption that Nichols’s sexual interactions with the two other inmates were nonconsensual would mean that Sheriff Duncan was on notice of a pattern of constitutional violations, rendering his failure to discipline or terminate Nichols deliberately indifferent. It’s of course possible the evidence here (or in some future case) could overcome the presumption. For instance, the majority opines that “[t]he evidence of Nichols’s conduct with one of the inmates” demonstrates no genuine dispute over consent “because the inmate testified that all the encounters were consensual.” Maj. Op. 17. In the limited context of this case, I don’t disagree. But taking a broader perspective, I find it troubling that this conclusion rests on uncritically accepting inmate testimony elicited approximately Page 27seven years after the sexual activity at issue and without considering the inmate’s possible motivations for denying any nonconsensual sex.

A presumption of nonconsent would allay this concern. Without one, prison administrators in this circuit have no incentive to conduct more reliable, contemporaneous investigations into whether sexual interaction between correctional officers and inmates is consensual. Neither state law nor the prison policy at issue provide such an incentive because neither turn on consent. See Walton , 135 F.4th at 1075 (noting national consensus that consent isn’t a defense to correctional officer engaging in sexual activity with a prisoner). So administrators like Sheriff Duncan can simply decline to investigate repeated incidents of sexual interactions, knowing that if the evidence is left unclear or undeveloped, this court will presume consent unless the inmate can retrospectively show “some form of coercion” under Graham . 4 741 F.3d at 1126. On the other hand, if we presume for purposes of Eighth Amendment sexual-assault claims that all sexual activity between inmates and correctional officers is nonconsensual, prison administrators will be encouraged to conduct contemporaneous investigations to marshal evidence on consent if they want to preserve their ability to avoid § 1983 liability for such claims. Contemporaneous investigations would obviously be more reliable and would have the added benefit of Page 28being conducted by prison officials in a better position to gather relevant materials than an inmate, including an inmate who is assaulted many years later.

The contemporary legal landscape and evolving standards of decency also support reconsidering Graham . “Supreme Court precedent is clear that bedrock Eighth Amendment principles require us to assess a prison official’s conduct against ‘the evolving standards of decency that mark the progress of a maturing society.’” Walton , 135 F.4th at 1074 (quoting Hudson v. McMillian , 503 U.S. 1, 8 (1992)); see also Crawford v. Cuomo , 796 F.3d 252, 259–60 (2d Cir. 2015) (relying on evolving standards of decency to adopt broad definition of sexual assault for Eighth Amendment claims, such that conduct permitted under 18-year-old precedent “would flunk its own test today”). We need not look far to find strong indications that our standard of decency has evolved in this context. As already noted, all 50 states and the District of Columbia prohibit sexual activity between inmates and correctional officers regardless of consent.⁵ See Walton , 135 F.4th at 1075. “Federal law similarly makes it a criminal offense for a federal law[-]enforcement officer to ‘knowingly engage in a sexual act’ with a prisoner, regardless of consent.” Id. at 1075 (cleaned up) (quoting 18 U.S.C. § 2243(b)). In fact, federal law has only become stricter over time. Since 1986, when § 2243(b) was first enacted, Congress has increased the statutory maximum sentence and expanded the definition of individuals in federal Page 29custody. See Violence Against Women and Department of Justice Reauthorization Act of 2005, Pub. L. No. 109-162, § 1177, 119 Stat. 2960, 3125; Adam Walsh Child Protection and Safety Act of 2006, Pub. L. No. 109-248, § 207, 120 Stat. 587, 615; Consolidated Appropriations Act, 2022, Pub. L. No. 117-103, § 1202, 136 Stat. 49, 923–25.

“[T]he takeaway is clear: our country’s legislatures have determined that sexual activity between prisoners and prison officials is a crime, intolerable conduct in a civilized society.” Walton , 135 F.4th at 1075. Moreover, this “aligned judgment of elected officials at the state and national level is consistent with Eighth Amendment jurisprudence recognizing the inherently vulnerable position of prisoners, especially in relation to prison officials.” Id. Indeed, in the years since Graham , the Sixth Circuit has joined the Ninth in presuming nonconsent in this context. See Hale , 18 F.4th at 854. And the Seventh Circuit, though it was not called upon to adopt the presumption in the case before it, saw “many reasons for why a presumption of nonconsent may prove to be the right answer to a difficult question.” 6 Walton , 135 F.4th at 1074.

So I urge this court to consider—albeit in some future case—whether a presumption of nonconsent is “the right answer.” Id. Without such a presumption, Graham ’s acknowledgment of “the power dynamics between prisoners and guardsPage 30 [that] make it difficult to discern consent from coercion” is mere lip service. 741 F.3d at 1126 (cleaned up) (quoting Wood , 692 F.3d at 1047). To truly recognize the inherently coercive nature of the prison environment, we should depart from Graham and adopt a rebuttable presumption of nonconsent for sexual interactions between correctional officers and inmates.