Uber had internal data identifying drivers with multiple harassment complaints as higher risk — and the company's response was to put them lower in the queue. Why are you allowing those people to drive at all? Lowering them in the queue is not the same as removing them.
This isn't my usual content — but the New York Times reporting on Uber's assault problem is fascinating and has been on my mind all week. I wanted to share my thoughts because some of the patterns in the story echo conversations we have constantly in medicine about institutional responsibility, patient (or user) safety, and how companies behave when problems are inconvenient.
If you haven't read the article or listened to the Daily podcast episode, it's worth doing both. Here's what struck me.
A New York Times reporter gained access to internal Uber records showing that the company has had a significant assault and harassment problem affecting:
This isn't a couple of isolated incidents. It's a longstanding pattern that Uber has known about internally for many years and has worked behind the scenes to manage.
The most striking part of the reporting, to me, was that Uber recognized this problem years ago and did internal analysis to identify situations associated with elevated risk.
Some of the patterns they flagged were predictable:
These are unsurprising. Most people's gut instinct would land on the same list.
They also identified patterns related to driver behavior history:
So they built an internal algorithm that placed those drivers lower in the queue when matching with potential passengers — essentially making it less likely (but not impossible) for high-risk drivers to be matched with high-risk passengers.
The algorithm response isn't nothing. But it raises an obvious follow-up:
Why are you allowing people to drive for your company at all if they have multiple complaints about harassment?
Lowering them in the queue is not the same as removing them. If a driver has accumulated multiple complaints from female passengers and Uber's response is to keep them on the platform but match them less often, that's a company decision that prioritizes driver retention over passenger safety.
This is the kind of pattern that comes up in medicine too. When a clinician accumulates patient complaints, malpractice cases, or board actions and the institutional response is to quietly limit their patient assignments rather than hold them accountable, we end up with the same problem: harm that's known internally and not visible to the public actually using the service.
Uber announced, around the time this story was being reported, that they were rolling out a feature allowing female passengers to choose female drivers.
But the internal documents tell a more interesting story.
According to the reporting, Uber had internally developed the female-driver-selection feature years before the announcement. They were prepared to roll it out.
Internal documents show that when Trump won in 2016, Uber's leadership decided the political climate wasn't right for rolling out a feature framed around protecting women from assault.
Take a moment to sit with that one.
The data showed female passengers faced real assault risk. The technical solution was already built. And the company decided that the political environment for "protecting women from assault" was unfavorable enough that they shelved the feature. For years.
Apparently the climate wasn't right for that — but the climate is always right for letting the problem continue.
The fact that they're rolling the feature out now is almost certainly a direct response to this NYT reporting rather than a genuine internal decision that women's safety became a priority.
There's a downstream issue worth noting: Uber has trouble recruiting female drivers. Why? Because of the same assault and harassment risk that the company has been internally aware of.
Female drivers don't want to be alone in cars with male passengers who may harass or assault them. Predictable. So:
The female-driver-selection feature, ironically, only helps if there are actually female drivers available — which there often aren't, because of the same underlying problem the feature is supposed to address.
The part of this story that struck me — in a "funny, not funny" way — was watching the response to the female driver feature.
There were men making angry videos about how it's "ridiculous" that women should be able to choose female drivers — and that they (the men) should similarly be able to choose male drivers.
I want to gently point out the disconnect there.
Female passengers choosing female drivers do it for one reason: safety from harassment and assault. It's a defensive choice. It's rooted in real risk that they have to manage in their daily life.
If a man demands the equivalent feature — the ability to select only male drivers — the underlying reason isn't the same. It's not because he's afraid of being harassed or assaulted by a female driver. It's because of sexism and misogyny.
These are not symmetric situations. The framing of "well, women have it, so men should too" treats them as if they're the same kind of preference. They are not.
The deeper thread in the reporting — and the one I keep thinking about — is this:
What kind of responsibility does a gig-working app have to keep its users safe?
Uber maintains, legally and operationally, that:
If the company won't take responsibility for the employment relationship with drivers, how can it justify taking responsibility for passenger safety when it's convenient? Or vice versa — if it's the entity controlling who can drive on the platform, how is it not responsible when those drivers harm passengers?
Same questions come up in medicine when we talk about hospital systems and the surgeons working under them, or about private equity-owned chop shops and the surgeons they employ. The structural answer companies want is "we're the platform, not the provider" — but the practical reality is that the platform shapes everything about how the service is delivered.
The Uber story is a great case study in a broader pattern:
This is the playbook. Once you can identify it, you start seeing it everywhere — in tech, in finance, in healthcare, in pharma, in plastic surgery influencer ecosystems.
The New York Times reporting on Uber's assault problem is excellent journalism that surfaces a pattern that's gone unaddressed for years. The reporting raises three things worth sitting with:
I'd strongly encourage you to read the NYT piece or listen to the Daily podcast episode. The reporting is great, and the questions it raises about institutional responsibility apply to a lot more than just rideshare.
What did you think of the story? Let me know.