Interviews with 20 people show Uber's 80-strong Special Investigations Unit, which deals with severe incidents, often puts Uber's interests ahead of riders'
Investigators say drivers get three “strikes” for bad behavior — and may just end up driving for Lyft. Tweets: @drewharwell , @elizabeth_joh , @goldfarb , and @laurastevens Tweets: Drew Harwell / @drewharwell : Huge story: Uber investigators said they were coached to protect the company and forbidden from alerting police to crimes. One driver, cleared by investigation, went on to allegedly rape a passenger https://www.washingtonpost.com/ ... @GregBensinger https://twitter.com/... Elizabeth Joh / @elizabeth_joh : This is *the* problem with internal/corporate policing—there is a fundamental conflict of interest. https://twitter.com/... Zachary Goldfarb / @goldfarb : Critical reading on a busy day: @GregBensinger's investigation into the Uber unit that looks at allegations of wrongdoing - and how it protects the company over riders https://wapo.st/2l956i8 Laura Stevens / @laurastevens : Uber has an entire unit for following up on rides gone bad. They say it leaves unfit drivers on the road. Via @GregBensinger https://www.washingtonpost.com/ ...
Context & Ripple Effects
This investigation closes the loop on a year of embedded reporting on Uber's incident-handling apparatus: earlier coverage went inside what was then a 60-plus person team handling roughly 1,200 severe reports a week and its expansion into a ~130-investigator unit, but treated it primarily as a workplace story about investigator burnout. Drew Harwell's interviews with 20 people reframe it as a governance story — investigators say they were coached to protect Uber and barred from alerting police to crimes.
The consequences were already visible in Uber's own disclosures months later: the company's [[a:948563|safety report admitted only 37% of the 464 US rape cases in 2017–2018 had been reported to law enforcement]], which Dara Khosrowshahi defended in an accompanying interview. And by 2022 The Markup found Uber still responding slowly to law enforcement requests, leaving drivers exposed to repeat attackers using fake accounts — suggesting the structural incentives this article surfaced persisted.
First-order effects
- Riders reporting severe incidents face an investigative process optimized for Uber's liability rather than their safety, and at least one driver cleared by the unit was later accused of raping a passenger.
- Drivers disciplined under the three-strike policy lose their Uber access but can keep driving — Lyft has no visibility into those strike records, so removal from one platform does not mean removal from rideshare.
Second-order effects
- Lyft inherits the disciplinary overflow: Uber's strikes effectively subsidize its rival's supply while exporting the safety risk, giving both platforms a shared interest in a cross-platform conduct database neither has built.
- Police and prosecutors who assumed Uber's internal unit functioned as a first responder now have documented grounds to demand direct notification protocols, shifting incident triage off Uber's own terms.
Third-order effects
- If self-policing keeps producing under-reported violent incidents, the likely endpoint is mandated external reporting of severe ride incidents — moving severe-case adjudication from corporate units like the SIU toward independent or statutory oversight, the outcome Elizabeth Joh flags as inherent in corporate policing.
The trend: Platform self-policing of severe safety incidents is heading from internal discretion toward externally mandated reporting and cross-platform accountability.