Self-driving cars have moved from science fiction to real trials on public roads. Supporters claim that they will save many lives and should therefore take over from human drivers quickly. I agree that this technology has great potential to improve safety, but I believe a rapid replacement of human drivers would be a mistake.
There is little doubt that most road accidents are caused by human error. Drivers become tired on long journeys, check their phones at traffic lights, misjudge speed or drive after drinking alcohol. A computer does none of these things. It can monitor every direction at once, react in a fraction of a second and obey speed limits without exception. In theory, then, a road network dominated by automated vehicles could see far fewer serious collisions, and emergency services and hospitals would benefit as a result.
However, the phrase "as soon as possible" ignores several practical problems. Current systems perform well on clear motorways but can still be confused by heavy snow, faded road markings or a police officer directing traffic by hand. Legal questions are also unresolved: if a driverless car injures a pedestrian, it is not clear whether the owner, the manufacturer or the software designer is responsible. Furthermore, many rural roads lack the detailed digital maps that these vehicles rely on. Forcing a fast transition would risk public trust, since a few well-publicised accidents could turn people against the technology altogether.
In conclusion, I accept that driverless cars may one day make roads considerably safer than they are now. Nevertheless, the change should be gradual, with careful testing, clear laws on liability and investment in infrastructure, rather than a rush to remove human drivers before the technology is proven in every situation.