Risks for Clicks: Why the Algorithm Keeps Rewarding Shock Content

A New York Times investigation traces how recommendation algorithms push creators toward increasingly dangerous stunts — from Trevor Jacob's plane bailout to the "door kick challenge." Here's what the incentive problem means for working creators.

By Board of Creators · 2026-09-15

Two videos, four years apart, tell the whole story. In the first, a creator named Trevor Jacob bails out of a small aircraft mid-flight and lets it crash — because the footage would perform. In the second, a livestreamer who called himself "Chud the Builder" turned provocation into a format, streaming confrontations until one ended with a man shot outside a Tennessee courthouse.

A new investigation by Mark Walker in The New York Times, published September 12, connects those dots. Its finding is blunt: a growing number of social media creators are pushing the boundaries of safety and good taste because the recommendation systems that decide who gets seen are built to reward exactly that.

For professional creators, this isn't an abstract ethics debate. The same algorithms that can build a career can also pull it toward ever-riskier content — and understanding that pull is now basic career hygiene.

What the Times found

The Times' reporting, filed from the communities where these stunts land, describes a clear pattern: creators discover that outrage, danger and humiliation outperform everything else, and the platforms' recommendation engines amplify whichever videos hold attention longest.

Trevor Jacob's plane-crash video is the archetype. Jacob recorded himself abandoning the aircraft mid-flight and posted the footage online, where it drew enormous attention — and eventually federal scrutiny, after investigators concluded the crash was staged for views.

The story also touches Loudoun County, Virginia, where the so-called "door kick challenge" — filming yourself kicking or pounding on strangers' front doors, then running — spread widely enough this spring to flood local 911 lines and force the county sheriff's office to issue public warnings that the prank could turn deadly.

The "Chud the Builder" case

The darkest example in the current news cycle is Dalton Eatherly, the 28-year-old streamer known as "Chud the Builder." Eatherly built an audience by streaming confrontational, racially charged encounters with strangers — a format that thrived on outrage clips shared across platforms.

In May, Eatherly was arrested in Clarksville, Tennessee, and charged with attempted murder after allegedly shooting and wounding a Black man outside the Montgomery County courthouse. A judge set his bond at $1 million. As the Associated Press reported, the case ignited a debate about whether "free speech" framing can launder what critics describe as race-baiting for engagement.

The victim, Army veteran Joshua Fox, later told Slate: "I just don't wanna be known as the Black guy Chud the Builder shot." The algorithms that rewarded the format moved on within days. The people it touched don't get to.

Why the algorithm tilts toward danger

None of this requires the platforms to intend harm. Recommendation systems on YouTube, TikTok, Instagram and Kick are optimized for watch time and session length — and emotionally extreme content simply holds attention better than measured content.

That creates a ratchet. A creator who gets a spike from a risky video learns the lesson instantly: do it again, bigger. Audiences habituate, so the next video has to go further to get the same response. Researchers call it the escalation trap, and it's a structural feature of engagement-based ranking, not a bug any single moderation policy can patch.

Kick, the streaming platform that positions itself as a looser alternative to Twitch, comes in for particular scrutiny in the current debate — its permissive moderation has made it a landing zone for creators whose formats got them suspended elsewhere.

What this means for working creators

If you make content for a living, the takeaway isn't "never be edgy." It's to know exactly where your floor is before the algorithm starts negotiating with you.

First, separate reach from reputation. Shock content can spike a channel and permanently cap its brand-deal ceiling in the same quarter. Sponsors increasingly screen for exactly this history.

Second, watch your own escalation curve. If each video needs to be louder than the last to hold your numbers, that's the ratchet — not organic growth.

Third, document context. Creators who do stunts, pranks or confrontation-adjacent content should keep releases, safety plans and location permits. When something goes wrong — and the Loudoun County 911 surge shows how fast it can — that paper trail is the difference between a misunderstanding and a charge.

Finally, remember that platform incentives change faster than platform values statements. Build formats you'd be proud of if the algorithm disappeared tomorrow.

The bottom line

The Times' reporting lands at a moment when regulators, advertisers and audiences are all asking the same question: who is responsible when the feed rewards harm? The honest answer for now is: mostly the creator in the frame.

Until ranking systems stop paying out for danger, the safest career strategy is the oldest one — make things people are glad they watched.

Source: The New York Times