The New York Times reported thatUber considered but rejected tools to make trips safer: matching algorithms to avoid matching high-risk riders with high-risk drivers, mandatory video recording, and pairing female passengers with female drivers. But the Times says Uber delayed or did not require adoption of some of the most promising programs, nor did it warn passengers about factors it linked to attacks.
The Times quotes an internal Uber brainstorming document: “Our purpose/goal is not to be the police. Our bar is much lower and our goal is to protect the company and set the tolerable risk level for our operations.”
The Times highlights a specific Uber algorithm called Safety Risk Assessed Dispatch, which evaluated the risks of potential pairings of drivers and passengers. Uber found that the system correctly anticipated 15% of sexual assaults, and an internal presentation called the tool “the most effective intervention for preventing sexual assaults.” But the system still dispatched rides deemed high risk.
As early as 2014, Uber considered using cameras to monitor drivers, remarking in internal discussions that the lack of video “leaves gaps in our safety ecosystem.” But the Times says Uber decided not to implement video monitoring because it was concerned that videos might prompt worker misclassification litigation — alleging that drivers are employees rather than drivers, and must receive employment benefits.
In Saudi Arabia, Uber launched a feature to pair female passengers only with female drivers. But the Times reports that Uber declined to offer that feature in the United States based on concerns about “culture wars, political blowback, and … gender discrimination and other lawsuits.”
Uber considered publishing information about factors that were correlated with sexual assault, such as rides late at night and in particular areas. But internal discussions flagged “serious business implications” of revealing this information, and Uber decided not to tell users.