The comparison sounds easier than it is
A fair safety comparison would ideally measure the same kinds of events over the same years, use comparable exposure data and cover a representative share of trips nationwide. That dataset does not currently exist for taxis and rideshare services.
The U.S. Government Accountability Office said in 2026 that available federal, local and company data did not provide a complete nationwide picture of assaults involving rideshare and taxi trips. That alone makes sweeping “safer than” claims difficult to defend.
Rideshare companies disclose more standardized national data — but with limits
Uber and Lyft publish safety reports covering defined categories of serious incidents. Those reports provide nationwide company data that are unusually detailed for transportation services. But they are company-produced, use specified categories and are released on reporting cycles that do not always provide current-year data.
GAO noted that comparable public assault counts for 2023 through 2025 were not available from selected rideshare companies as of June 2026. So even within rideshare, the latest complete public picture can lag behind the current date.
Taxi data are scattered across local systems
Taxi oversight has historically been local. Cities may license drivers, inspect vehicles, collect complaints or regulate dispatch companies, while another jurisdiction uses a different structure. A national taxi company can provide its own incident data, but that does not represent the entire U.S. taxi industry.
GAO obtained nonpublic assault information from only a small number of taxi companies and expressly cautioned that the information was not nationally representative. One national company reported a limited number of verified physical and sexual assaults from 2021 through 2025, but that is not a nationwide taxi rate.
Police records do not neatly solve the problem
Police data can contain assaults involving drivers or passengers, but transportation mode may not be captured in a standardized field. An officer’s narrative might mention Uber, Lyft, rideshare, cab or taxi without placing the incident into a consistent national category.
Local crime records are therefore useful for city-level research and incident reconstruction, but combining them into a national mode-to-mode safety rate requires assumptions the source data were not built to support.
Crash safety is a separate question from assault safety
Another source of confusion is the word “safe.” Assault risk, fatal-crash risk, injury-crash risk, driver screening, vehicle condition and accessibility are different dimensions. A transportation mode could perform better on one measure and worse on another.
Research on rideshare availability has even produced findings that appear to point in different directions depending on the outcome measured: some studies associate rideshare availability with fewer alcohol-related fatalities, while local studies have found increased injury-crash exposure with greater rideshare activity. Those findings are not a taxi comparison and should not be turned into one.
The better answer is transparent about the limits
A responsible taxi-versus-rideshare article can still be useful. It can compare screening systems in a specific city, explain vehicle and insurance rules, identify available assault or crash datasets and describe what each source actually measures.
What it should not do is turn fragmented numbers into a national ranking that the evidence cannot support. As of 2026, the honest answer remains that the United States lacks one unified dataset capable of declaring taxis, Uber or Lyft categorically safer nationwide.
Sources
