Twenty seven minutes. That is how long passed between the first message Braden Markus received and his death. He was fifteen, and he played football.
Jordan DeMay was sixteen. The first message arrived at 10.19pm. He was dead by around 3.45am, roughly five hours later. In the cases documented in ESPN’s reporting on the FBI’s warning to young athletes, every known victim who died by suicide did so within twenty four hours of first contact, and most within hours. The men responsible in the DeMay case, Samuel and Samson Ogoshi, were each sentenced to seventeen and a half years in September 2024.
This is the offence the FBI and the NCAA announced a joint programme against on 10 August 2026. The Bureau’s own release is explicit that the purpose is awareness rather than investigation: "Our goal is to raise awareness so students can protect themselves and prevent further victimisation." It asks coaches, compliance staff and athletic department leadership to pass the information on and to act as the first trusted adults an athlete can turn to.
Jose Perez, an FBI Operations Director, framed it as reaching students before an incident as well as after: "This initiative is about making sure student-athletes, and anyone who may become a target, know they are not alone and that help is available." Jon Duncan, the NCAA’s Vice President for Enforcement, said the aim was to "break down barriers so law enforcement can help prevent further abuse and connect college athletes who may be targeted with the right support."
It is worth noting what the release is not. There is no memorandum of understanding, no data-sharing mechanism, no joint taskforce and no funding commitment. It is an awareness and distribution partnership. The tactics it describes are credential-focused: unsolicited messages claiming an account will be disabled without a verification code, and emails referencing a new login with a link to reset a password. Reporting routes given are ncii.ic3.gov, tips.fbi.gov and 1-800-CALL-FBI.
The scale is in the FBI’s own reporting. Its Internet Crime Complaint Center received more than 75,000 sextortion-related submissions during 2025, of which more than 5,700 were referred to the National Center for Missing and Exploited Children as involving minors. An earlier national alert covering October 2021 to March 2023 recorded over 13,000 reports of financially motivated sextortion of minors and at least 12,600 victims, most of them boys.
An offence measured in hours
Almost every safeguarding process in sport is built for a slower crime. Reports are triaged, investigations are opened, welfare teams convene. That architecture assumes there is time.
Here there is not. A child can be approached, groomed, extorted and lost between dinner and breakfast, without anyone in the building knowing a conversation took place.
Australian data describes the same compressed window. The Australian Institute of Criminology surveyed 1,953 Australians aged 16 to 18 in mid-2024 and found that among those who had been sexually extorted, 45.9% were threatened less than a day after first contact and 58.6% within a week. Half of those who met the offender online were asked to move the conversation to another platform, most often Snapchat, Instagram or WhatsApp.
ESPN’s reporting adds detail the FBI release does not carry. Speaking to ESPN, the Bureau described a shift in who is targeted, from predominantly female victims toward males aged fourteen to seventeen, and its Cyber Division noted that student-athletes maintain large public-facing followings, in many cases expanded further by name, image and likeness activity. Its child exploitation unit put the exposure plainly: school athletes have a larger online footprint, and they have more to lose.
Every documented case begins in private
Read the case reporting and a consistent structure appears. The offender builds a plausible profile, often using stolen photographs, follows the target and their friends so the account looks legitimate, opens with something unremarkable, establishes rapport, and moves to a request for an image. The moment an image arrives, the approach turns to extortion. Instagram is the platform named most often in the reporting.
All of that sequence happens in private, in a direct message between two accounts, in a conversation nobody else is party to. There is no post to flag, no comment to hide, no hashtag to track. By the time anything about it is visible externally, the harm has already occurred.
The measurement problem the sector has documented itself
The sport sector has invested seriously in monitoring online abuse, and the published work is careful about what it covers.
World Athletics published a four-year longitudinal analysis in December 2024 spanning Tokyo, Oregon, Budapest and Paris, examining more than 1.4 million posts and comments across 2,438 athletes. It is the most substantial dataset of its kind. It also states its own boundary explicitly: the study "excluded direct messaging or non-targeted abuse conversation; where the athletes were mentioned but not directly tagged, suggesting the issue is even greater than the study highlights".
The International Olympic Committee published its Online Abuse and Sport report on 7 August 2026, drawing on the AI monitoring it ran at Paris 2024 and Milano Cortina 2026. Across both Games it analysed some 4.3 million posts and comments and verified around 25,000 abusive messages. Its Paris figures were 10,200 verified abusive messages from roughly 20,000 monitored accounts; at Milano Cortina, 14,500 from just over 6,000 accounts. On the IOC’s own numbers that is a rise of about 42% in verified abuse against roughly a third as many monitored accounts. Every public description of that service refers to posts and comments.
Both bodies deserve credit here. Each is measuring a real harm rigorously, and World Athletics went to the trouble of stating the limitation of its own dataset rather than leaving a reader to infer it. The point is narrower and more uncomfortable. The surface the FBI and the NCAA have just prioritised is the one the sector’s measurement infrastructure was never built to see.
What sits alongside it
Direct messages are not only a sextortion vector. The NCAA published research in November 2025 finding that 36% of Division I men’s basketball players had received betting-related abuse on social media in the previous year, and that in a survey of 7,493 Division I athletes, 48% reported facing online abuse of some kind. Where a betting market exists there is a motivated, identifiable counterparty, and a private channel to reach the athlete through.
The pattern holds across the categories. Public abuse is loud, measurable and increasingly well covered. Private abuse is targeted, escalates quickly, and is where the conduct that ends careers and lives actually takes place.
What detection in the private layer requires
Guardii was built for this surface. Three properties matter, and they are properties of the architecture rather than the feature list.
Detection happens before the athlete reads the message. Abusive or predatory inbound content is identified and suppressed at the point of arrival, so protection does not depend on the target absorbing the abuse first. Set that against Braden Markus’s twenty seven minutes. A model that starts only once the athlete has read the message, understood what is happening, and decided to tell someone is not operating on the same clock as the offender.
Reporting does not require re-exposure. Where a person must re-read what was sent in order to escalate it, the reporting rate reflects what victims can bear rather than what happened. The record is assembled from the incident log, at a time of the athlete’s choosing.
The record is built to survive challenge. Content is committed in a form that establishes what was sent, when, and by whom, without requiring anyone to read it in order to hold it. That is a stronger privacy position and a stronger evidentiary position at once, and for a governing body facing a duty of care question it is the difference between an allegation and a record.
The FBI and the NCAA have now told the sport sector where this harm lives. Past the timeline, past the replies, past the hashtag. It lives in the inbox, and it moves fast.