In June 2016, a 28-year-old man named Antonio Perkins was shot and killed in Chicago while his phone broadcast everything live to Facebook. It was one of the first deaths captured in real time on a mainstream social media platform — and it forced a question the industry still hasn’t fully answered: what responsibility do platforms carry when tragedy goes live to a global audience?
Since that case, dozens of deaths have been broadcast in real time across Facebook Live, Instagram Live, YouTube, and smaller platforms. Some involved murder victims. Some involved accidents. Some involved suicide. Each one exposed the gap between how fast live content spreads and how slowly platforms respond.
This article documents the most widely-reported and independently verified cases, year by year. It also covers how platforms responded, what laws changed, and what these incidents reveal about the intersection of social media, public safety, and accountability. All cases are drawn from reporting by established news organizations including the BBC, The Guardian, The Washington Post, Reuters, and the Associated Press.
What Qualifies as a Documented Livestream Death?
A livestream death occurs when a person’s death — or the direct circumstances immediately causing it — is broadcast in real time to a live audience on a streaming platform. The word live is the key distinction here.
Three scenarios fall under this definition:
- Death during the person’s own stream: The individual streaming is killed or dies during the broadcast — through a shooting, accident, medical event, or suicide.
- Death captured incidentally: A bystander is streaming unrelated content when a death occurs in frame.
- Death broadcast by a perpetrator: A killer livestreams or immediately posts video of a murder as it happens or immediately after.
Cases where footage was recorded but not streamed live — dash-cam video, CCTV, phone recordings — do not meet this definition. That distinction matters practically: live content creates a window in which platforms have no editing or pre-upload review capability. Once a stream is live, the content reaches viewers before any moderation system can evaluate it. This is the core problem that makes livestream deaths uniquely challenging for both platforms and policymakers.
Complete Timeline of Documented Livestream Deaths (2016–2024)
2016 — The Year the Problem Became Visible
Facebook opened Facebook Live to all users in April 2016. Within months, three internationally-covered fatal incidents had occurred on the platform. Each revealed a different dimension of the same problem.
Antonio Perkins — June 21, 2016 | Facebook Live | Chicago, USA Perkins, 28, was shot and killed while streaming in his neighborhood. The broadcast captured the shooting in real time before the feed cut out. He had been talking with friends on a residential street when gunfire erupted. The incident was among the first to raise public questions about whether any mechanism existed to detect or respond to violence in live social media content.
Philando Castile Aftermath — July 6, 2016 | Facebook Live | Falcon Heights, Minnesota, USA Castile was shot during a traffic stop by police officer Jeronimo Yanez. His girlfriend, Diamond Reynolds, began livestreaming on her phone in the immediate aftermath, while Castile remained in the car gravely wounded. He died shortly after. The footage — which also showed Reynolds’s four-year-old daughter in the backseat — was viewed more than 2.5 million times within 24 hours. Facebook briefly removed it, then restored it with a content warning. The case became a defining moment for the Black Lives Matter movement and a landmark example of social media’s role in documenting police conduct.
Katelyn Nicole Davis — December 30, 2016 | Facebook Live | Cedartown, Georgia, USA A 12-year-old girl died by suicide during a Facebook Live broadcast. The video remained accessible for weeks despite multiple user reports — a severe failure in Facebook’s content moderation response time, particularly involving a minor. In line with responsible reporting standards, no method details are included in this article. The case became central to later debates about platform duty of care toward minors and the adequacy of human review timelines for live content.
2017 — A Wave of High-Profile Cases
Lavontay White Jr. — February 2017 | Facebook Live | Chicago, USA Two-year-old Lavontay White Jr. was shot and killed while his uncle was driving and streaming live on Facebook. The car was struck by gunfire; the uncle survived, the toddler did not. The video circulated on social media before Facebook removed it. Lavontay’s death added child casualties to the growing record of live-streamed gun violence in Chicago.
Robert Godwin Sr. — April 16, 2017 | Facebook | Cleveland, Ohio, USA On Easter Sunday, 74-year-old Robert Godwin Sr. was shot and killed on a public sidewalk by a man named Steve Stephens. Stephens posted the video of the murder to Facebook shortly after it occurred. He had also announced his intentions on Facebook Live before the killing. Facebook removed the video within roughly two hours of the first report — a timeline widely criticized as inadequate. Stephens was located in Pennsylvania two days later and died by suicide during a police chase. Facebook CEO Mark Zuckerberg acknowledged the incident publicly and committed to improving reporting tools and moderation speed.
Wuttisan Wongtalay and His Daughter — April 25, 2017 | Facebook Live | Phuket, Thailand A man livestreamed the murder of his 11-month-old daughter before taking his own life. The videos remained on Facebook for approximately 24 hours before being removed. Zuckerberg addressed this incident specifically in a public statement, calling it a failure and announcing a review of the company’s content moderation systems. The case drew international attention to Facebook’s cross-border moderation capacity and the speed gap between violent live content and platform response.
Jacqueline Sanchez — July 22, 2017 | Instagram Live | Stockton, California, USA Eighteen-year-old Obdulia Sanchez was driving under the influence while streaming on Instagram Live when her car left the road and crashed. Her 14-year-old sister Jacqueline was ejected from the vehicle and died at the scene. The broadcast captured the crash and its immediate aftermath. Obdulia was convicted of gross vehicular manslaughter and sentenced to six years in prison. The case became a frequently-cited example in road safety campaigns and legislative discussions about distracted and impaired driving. It also highlighted the role of Instagram Live in contexts outside the violence-centric debates of the same period.
2019 — The Defining Case
Christchurch Mosque Shootings — March 15, 2019 | Facebook Live | Christchurch, New Zealand This is the most extensively documented and consequential livestreamed mass violence event on record. Brenton Harrison Tarrant livestreamed his attacks on Al Noor Mosque and Linwood Islamic Centre for approximately 17 minutes on Facebook Live. Fifty-one people were killed and 40 were injured. Facebook’s own data showed that no user reported the video to the platform during the live broadcast. The footage circulated across multiple platforms in the hours that followed — Facebook alone attempted to remove 1.5 million copies in the first 24 hours, blocking 1.2 million at upload. The attack and its viral spread led directly to the Christchurch Call to Action, the most significant international policy response to online violent content since platforms began to scale.
2020–2024 — Ongoing, Expanding Volume
The volume of livestreaming globally has grown dramatically since 2019. Fatal incidents captured on live platforms have continued to occur in the US and internationally — including shootings documented on Instagram Live and Facebook Live, as well as health emergencies during extended gaming or streaming sessions. Detailed Western press documentation of cases during this period has been uneven, and several incidents have been verified locally but not confirmed by multiple international news organizations.
The trend is clear: as livestreaming becomes a default feature across every major platform, the incidence of fatal events being captured live will continue to grow in proportion to usage.
Summary Table: Major Documented Livestream Deaths by Year
| Year | Victim / Incident | Platform | Type | Reported Outcome |
|---|---|---|---|---|
| Jun 2016 | Antonio Perkins | Facebook Live | Shooting | Fatal |
| Jul 2016 | Philando Castile (aftermath streamed) | Facebook Live | Police shooting | Fatal |
| Dec 2016 | Katelyn Nicole Davis | Facebook Live | Suicide | Fatal |
| Feb 2017 | Lavontay White Jr. | Facebook Live | Drive-by shooting | Fatal (age 2) |
| Apr 2017 | Robert Godwin Sr. | Murder | Fatal (age 74) | |
| Apr 2017 | Wuttisan Wongtalay’s daughter | Facebook Live | Murder-suicide | Fatal (age 11 months) |
| Jul 2017 | Jacqueline Sanchez | Instagram Live | Impaired driving crash | Fatal (age 14) |
| Mar 2019 | Christchurch mosque attack (51 victims) | Facebook Live | Mass shooting | 51 Fatal, 40 Injured |
All cases verified by multiple major international news organizations. List is not exhaustive.
How Social Platforms Responded to On-Stream Deaths
Each platform’s response evolved under pressure from public criticism, government scrutiny, and internal review. Here is how the major players shifted their approach between 2016 and 2024.
Facebook / Meta Facebook’s response changed in identifiable stages. Before 2017, removal of violent live content was almost entirely reactive — waiting for user reports before taking action. After the Godwin murder in April 2017, Facebook announced it would hire 3,000 additional content moderators globally. After the Christchurch attack in 2019, it implemented a live streaming ban for repeat policy violators and began deploying AI classifiers designed to detect violent imagery during live broadcasts — not just after.
Facebook also introduced a rolling delay feature for users who have previously violated Live policies: offenders are prevented from using Facebook Live for a set period. This remains the only consumer-facing consequence mechanism Facebook has applied at scale.
Instagram Following the Sanchez case in 2017, Instagram accelerated the rollout of automated flagging for potentially dangerous content during live sessions. The platform began surfacing mental health resources to users who searched for terms associated with self-harm and began restricting certain Live features for underage accounts.
YouTube Live YouTube’s primary response to livestreamed violence has centered on expanding its use of GIFCT (the Global Internet Forum to Counter Terrorism) — a cross-platform hash-matching database that allows known violent content hashes to be shared between major platforms. After Christchurch, YouTube committed to more aggressive real-time application of GIFCT tools during live streams.
Twitch Twitch’s documented deaths during streams have primarily involved health emergencies — cardiac events and strokes during marathon gaming sessions — rather than violence. The platform has no specific policy requiring removal of deceased streamers’ archives without a family request, which has led to posthumous content remaining indefinitely accessible in several cases. Twitch added a dedicated “Safety Advisory Council” in 2020 partly to address these grey areas.
The Universal Gap Across every platform, in every documented case, the same structural gap appears: the time between a live fatal event and the first human report to the platform is longer than the event itself. The Christchurch data — zero reports during 17 minutes of live broadcast — is the starkest illustration. The implication is that human reporting cannot be the primary defense for live content. Algorithmic real-time detection must be, yet no platform has demonstrated reliable detection at scale without significant false positives.
What These Cases Changed in Law and Digital Safety Policy
The cumulative record of livestream deaths has produced measurable policy change, even if the pace has frustrated advocates.
New Zealand’s Legislative Response (2019–2021) Following Christchurch, New Zealand’s Chief Censor classified the attacker’s manifesto as objectionable under existing law, making its distribution a criminal offense. Parliament subsequently passed the Films, Videos, and Publications Classification (Urgent Interim Classification of Publications and Prevention of Online Harm) Amendment Act 2021, directly expanding the legal framework to address online extremist content distribution.
The Christchurch Call (2019) Forty-eight governments and all major social media platforms signed this joint commitment to eliminate terrorist and violent extremist content online. The Call established shared reporting protocols, expanded cross-platform hash-sharing requirements, and created direct communication channels between governments and platforms that had not previously existed. It remains the most concrete international policy outcome directly traceable to a single livestreamed event.
Australia’s Online Safety Act (2021) Australia’s revised Online Safety Act explicitly includes live streaming in its regulatory scope. Platforms operating in Australia are legally required to remove harmful live content within tight timeframes or face substantial financial penalties. This placed Australia among the first jurisdictions to legislate specifically around live streaming as a distinct content category.
US Section 230 and Platform Liability Debates In the US, livestream deaths have become a recurring reference point in congressional debates about whether Section 230 of the Communications Decency Act — which broadly shields platforms from liability for user-generated content — adequately accounts for algorithmic amplification of harmful content. No comprehensive federal legislation had passed directly addressing livestreamed violence as of early 2024, but multiple bills have been introduced in each congressional session since 2017.
Common Myths About Livestreamed Deaths
Myth 1: Most livestream deaths are staged for attention The documented record does not support this. The majority of verified cases involve unplanned violence — shootings and accidents — that occurred during an active broadcast. A smaller subset involves perpetrators who deliberately used livestreaming as part of an attack, as in the Christchurch case. But even there, the deaths were not staged.
Myth 2: Platforms can detect and stop violent live content in real time Current AI moderation tools improve yearly but remain imperfect in live contexts. Real-time detection must balance false positives — flagging innocent content, disrupting legitimate streams — against detection rates. The Christchurch attack evaded real-time algorithmic detection entirely during its 17-minute broadcast.
Myth 3: Removing the video removes the harm Once a stream reaches even a small live audience, removal cannot undo the psychological impact on those viewers or prevent further copying. Research from the Oxford Internet Institute has noted that the social amplification of violent content — sharing, reacting, discussing — extends harm well beyond the original broadcast window.
Myth 4: Watching these videos is a victimless act of information-gathering Trauma psychology research consistently finds that repeated exposure to graphic death footage can produce symptoms closely resembling PTSD. This has been documented among first responders, content moderators, and journalists. Viewing such material once for informational purposes carries lower risk than seeking it out repeatedly, but the risk is not zero.
Frequently Asked Questions
What was the first major death captured on a social media livestream? The shooting death of Antonio Perkins in Chicago in June 2016 is widely considered the first widely-covered fatal incident on a mainstream social media livestream. It occurred within months of Facebook opening Live to all users and was the first case to generate national media coverage focused on platform accountability — as opposed to earlier, smaller-platform incidents that received limited attention.
How quickly are violent livestreams typically removed after being reported? Removal times vary by platform and incident. Facebook has stated that most reported content is reviewed within 24 hours. However, live content presents a categorically different challenge: by the time a report is submitted and reviewed, a fatal event may be long over and the content already shared. The Christchurch attack ran for 17 minutes live with zero user reports during the broadcast. The video then circulated for hours before systemic removal.
Are social media platforms legally responsible for deaths shown in livestreams? In most jurisdictions, no. US Section 230 of the Communications Decency Act broadly shields platforms from liability for user-generated content. However, this protection is not absolute — several legal challenges have argued that algorithmic amplification of harmful content creates liability distinct from passive hosting. Courts in Australia, New Zealand, and the EU have taken different positions. No major platform has been held criminally liable for a livestreamed death as of 2024.
Why do perpetrators use livestreams during violent events? Research on the Christchurch attacker and similar cases suggests that livestreaming serves as both a mechanism for immediate audience reach and a form of psychological validation. Criminologists studying this pattern — sometimes called “performance violence” — note that the live audience, however small, functions as a component of the act itself. This makes livestream capability, in extreme cases, a factor that can shape violent behavior rather than merely document it.
Has anyone been criminally charged specifically for livestreaming a death? Yes. Obdulia Sanchez was convicted of gross vehicular manslaughter in California in connection with the Instagram Live crash that killed her sister. In the Godwin case, Steve Stephens was charged with aggravated murder before his death. Brenton Tarrant was convicted on 51 counts of murder and 40 counts of attempted murder in New Zealand and sentenced to life imprisonment without parole — the first such sentence in New Zealand history.
How do gaming-related stream deaths differ from violence-related cases? Gaming-related stream deaths — primarily cardiac events and strokes documented during marathon streaming sessions, with cases reported in South Korea, Taiwan, and China — are distinct in cause but raise similar platform questions about posthumous content management. Most platforms have no automatic process for identifying or handling a deceased streamer’s archive without a family member actively requesting removal.
Where can someone find support after watching disturbing livestream content? Several organizations provide support for trauma from online content exposure. The Crisis Text Line (text HOME to 741741 in the US, UK, Canada, and Ireland) serves people in acute distress. The Dart Center for Journalism and Trauma (dartcenter.org) specifically supports journalists and researchers who work with graphic content professionally. International crisis resources are available at findahelpline.com.
Conclusion
Documented livestream deaths represent one of the most uncomfortable and consequential challenges social media platforms have faced in the past decade. From Chicago in 2016 to Christchurch in 2019, each incident stripped away the idea that a livestreaming platform is a neutral tool — and forced the question of what obligations come with building infrastructure that lets any event reach millions of people before anyone can stop it.
The cases in this article collectively drove billions of dollars in content moderation investment, new legislation across at least three continents, and the first significant international policy agreements on violent online content. They also left behind families who lost people, viewers who cannot unsee what they witnessed, and a regulatory landscape that is still catching up with the technology.
For journalists, researchers, and policymakers, the most useful thing this record provides is specificity: not “social media can be dangerous” as an abstraction, but dates, names, platforms, response times, and outcomes. That specificity is what makes accountability possible.
Crisis and Support Resources
- US Suicide & Crisis Lifeline: Call or text 988
- Crisis Text Line: Text HOME to 741741 (US, UK, Canada, Ireland)
- International resources: findahelpline.com
- Trauma support for journalists/researchers: dartcenter.org
- Platform reporting: Each major platform maintains a mechanism to report harmful live content. Use it during the event — reports during a live broadcast are the most actionable.
This article is an independent editorial resource. All case summaries are drawn exclusively from reporting by established news organizations. No graphic content is linked or reproduced. Information accurate as of early 2024; some legal proceedings referenced may have developed further.
Your reading journey starts here—pick from our most loved content pieces.
