The City of Cape Town is showing a strong stance against distracted driving. The City has intentions to roll out AI-powered cameras that are capable of detecting if motorists are texting while driving or if they aren’t wearing their seatbelts.
While this technology has already been trialled in Cape Town, resulting in fines that were eventually overturned in court, the metro is now seeking formal approval from the National Director of Public Prosecutions to use the system permanently.
For motorists in Durban, these developments represent a significant shift in how road regulations might be enforced in major metropolitan areas in the future. The cameras move beyond traditional speed monitoring, using advanced software to peer inside vehicles and detect specific safety violations that were previously difficult to monitor without physical roadside stops.
What are the legal implications of using artificial intelligence for fines?
The transition to automated detection raises several legal questions. Technology law experts are closely examining the validity of AI-generated identifications, reports My Broadband. Dr Andrew Rens, a technology law consultant at Research ICT Africa, suggests that a primary concern is whether an identification made by an AI system is sufficient evidence on its own to secure a conviction.
Currently, South African law does not contain specific provisions that would allow a court to accept an AI conclusion as absolute proof of an offence.
The prosecution would need to go to great lengths to establish that the technology is reliable enough to serve as the sole basis for a fine. There is also the question of whether it is acceptable for a public body to use this specific type of technology for detection. Dr Rens clarifies that these systems use image recognition technology rather than generating new images. This means the software analyses an existing photograph to find patterns, such as a hand holding a mobile device.
How is image recognition technology used as evidence?
In order for a prosecution to be legally sound, an original, unaltered digital image must exist for all parties to examine. The artificial intelligence acts as a trigger by flagging a potential violation; however, the actual evidence remains the photograph itself. This allows a driver or a judge to verify whether the AI correctly identified a mobile phone or a seatbelt.
This process is somewhat similar to the current system of speed enforcement. Motorists often have to trust the accuracy of the equipment used to record their speed, as they cannot verify it themselves after the fact. However, systems are not error-free; there have been instances where number plate recognition has incorrectly identified a vehicle’s make or model.
While there is no specific legislation governing AI in South African traffic law yet, an analysis of a digital image remains a valid form of evidence that can lead to prosecution.
Why is there a push for more permanent cameras?
The number of fixed enforcement cameras has seen a massive increase in recent years. At the end of 2020, there were only 26 such cameras in operation in the Cape Town metro. By 2025, that figure had grown to 77 units. Despite this growth, safety officials argue that the current number is still far below the density seen in other major cities globally.
Alderman JP Smith, the Cape Town MEC for Safety and Security, confirmed in late 2025 that the existing network includes speed enforcement and red-light cameras. The newer models being introduced are more modern and less conspicuous. They utilise optical systems that do not require a visible flash, instead using infrared technology that is invisible to the human eye. This allows the cameras to operate effectively without alerting the driver with a bright light.
Do these cameras have a lasting impact on driver behaviour?
When considering whether these high-tech tools actually improve road safety, the results are mixed. Smith has noted that the impact of a camera on driver behaviour is generally limited to its immediate vicinity. Most drivers will slow down as they approach a known camera location and then speed up again once they are clear of it. Due to this, cameras are typically placed in accident hotspots or areas where the community has petitioned for red-light enforcement.
A major challenge with camera-based enforcement is the delay in consequences. Since fines often arrive a long time after the actual offence took place, there is a disconnect between the illegal behaviour and the punishment. This delay means that cameras alone may not be the most effective way to change a driver’s mindset over the long term. As these legal and technical battles continue, the outcome will likely determine if Durban road users will eventually encounter similar high-tech surveillance on their own commutes.
HOW TO LISTEN TO EAST COAST RADIO
- Listen to East Coast Radio on the FM (frequency modulation) spectrum between 94 and 95 FM on your radio.
- Listen live to ECR by clicking here or download the ECR App (iOS/Android).
- Listen to East Coast Radio on the DStv audio bouquet, channel 836.
- Switch to the audio bouquet on your Openview decoder and browse to channel 606
- Listen to us on Amazon Alexa.
Follow us on social media:
Image courtesy of iStock
