.
Image: Pexels
Ride-hailing companies increasingly pitch technology as the answer to safety problems that have long plagued taxi-like services worldwide.
In the past two weeks, two of South Africa’s biggest platforms, Bolt and Uber, pushed new safety-focused features into their South African apps: Bolt now requires riders to take a live selfie for profile verification, and Uber rolled out a Women Rider Preference (women-only matching) as part of its broader safety toolkit.
Both moves respond to public alarm and regulator pressure, but they also reveal the limits of what app-level tech can accomplish without better systems, policing and public infrastructure.
Bolt:
Bolt’s recently announced rollout in South Africa asks passengers to capture a live selfie to add to their Bolt profile.
The company says this helps drivers match a waiting rider to the booking and reduces fraud, while building a persistent identity signal for accountability.
Bolt already offers an in-app Emergency Assist button, audio trip recording (where available), driver identity verification via selfies, private phone-number masking and trip sharing.
Uber:
Uber’s new option lets female (and non-binary) riders request, and female drivers elect to accept, rides only with other women, a preference Uber has piloted in other markets.
The functionality sits on top of Uber’s Safety Toolkit features that include Share My Trip / Follow My Ride, RideCheck (automated alerts for long stops or potential crashes), an in-app emergency button connected to local services, phone-number anonymisation and the option for drivers to register dashcams.
Uber emphasises that Women's Preferences are one element of a broader safety strategy that also includes community funding and partnerships.
What about the other platforms?
Alternatives operating in South Africa, such as inDrive, highlight similar baseline protections.
This includes:
But the specific mix and maturity of features vary by operator and market.
Both Bolt’s selfie requirement and Uber’s women-matching feature are useful but narrow.
Bolt’s rider selfie addresses two real problems: mismatches at pickup points (drivers picking the wrong person), and anonymous accounts being used for fraud or to evade accountability.
Adding a persistent, app-linked photo can make a driver’s job easier and produce better evidence in a complaint. But a selfie is brittle: it’s trivial to spoof or share, and it does not prevent a user with malicious intent from creating multiple accounts or faking photos without robust liveness checks and backend fraud detection.
For drivers facing violent crime or carjacking, a picture is cold comfort compared with a fast emergency response.
Real-world risk remains high
One Bolt driver, who asked to remain anonymous for safety reasons, told Fast Company that theft is his single biggest threat on the job.
“Three times this year, I’ve had riders hold me up and steal my phone,” he said. “I don’t trust anyone anymore.”
He now screens every rider manually before accepting a trip, checking how long they’ve been on the platform and how many rides they’ve completed.
“If it’s less than 500 rides, it’s a no-go,” he explained. He’s also stopped mounting his phone on the dashboard, once a standard way to navigate, and now keeps it on his lap so it’s harder to grab.
His coping strategy shows how gaps in passenger verification and real-time protection push safety burdens onto drivers themselves.
Returns choice to women
Uber’s Women Rider Preference returns choice to women and can make everyday trips feel safer, particularly at night.
But it’s a matchmaking tweak; it does not change structural risks like route isolation, delayed emergency response, intimidation after the trip, or assaults that occur despite driver identity.
Uber’s feature is most protective where there are enough women drivers in a local supply pool; in many neighbourhoods, the skew of drivers may limit availability or increase wait times.
Both companies still rely heavily on reactive tools
Emergency Assist/SOS buttons, Share My Trip, anonymised numbers, and 24/7 safety teams are essential.
But they are mostly reactive, useful once something goes wrong and depend on the user to press a button, or on post-trip investigations to deliver justice.
Real-time prevention (detecting escalating risk before an incident) is harder and less present in current consumer feature sets.
Autonomous detection & machine learning
Uber has invested in automated signals (RideCheck-style detection of abnormal stops, route deviations and crashes) and uses Machine Learning (ML) for “safety data” to flag risky trips or accounts.
Those systems can surface anomalous trips for human review and sometimes preempt incidents.
Audio recording, dashcams and verification layers.
Bolt’s audio trip recording (offered in South Africa) and Uber’s dashcam integrations create evidence trails that discourage misconduct and help investigations when incidents occur, provided privacy and legal frameworks are respected.
Stronger, multi-factor identity verification (with privacy guardrails). Selfies are a start, but ride-hailing companies need to combine them with liveness detection, occasional random re-checks, and cross-checks to phone number and identity documents where privacy and law permit. Keep data retention minimal and transparent. (Bolt already uses selfies for drivers; extending robust liveness for riders helps.)
Faster, automated emergency escalation tied to first responders. In-app SOS should optionally push verified location + trip metadata directly to emergency services and local dispatch centres (not only to an internal safety team). That requires local partnerships and standards, but can cut seconds and save lives.
Real-time risk scoring and preemptive routing. These companies can use ML to build a live risk score for each trip (time of day, pickup location, account flags, driver history, route deviation patterns). If a trip crosses a risk threshold, the app could route drivers to safer pickup spots, alert the rider with safety tips, or temporarily require a second-factor confirmation (PIN) at pickup. Uber’s use of safety data and RideCheck indicates the technical path; it should be expanded.
Covert panic codes / silent alerts for drivers. Beyond a visible SOS button, drivers need a discreet way to signal duress (hidden tap sequences, voice triggers) that still shares real-time telemetry with the platform and local authorities.
Verified dashcams with privacy filters. It could be wise to allow drivers to register non-audio dashcams with automated upload on incident (video redacted for bystanders), with strict access policies for law enforcement and claim resolution. Uber has pilots for dashcam integration already.
Every added surveillance or verification step raises privacy concerns and the risk of exclusion. Over-aggressive identity checks can lock out low-income users without ID documents, and constant audio/video monitoring can chill vulnerable passengers or drivers.
Any deployment must balance safety gains with data minimisation, consent, and appeal mechanisms.