Patrick Magee (Has big tech backfired on robotaxis?, July 20) has missed the key point about relying on driver aids to be developed through to full driverlessness. The problem is that the lengthy intermediate stage creates unacceptable risk. As drivers become more reliant on aids, they become deskilled and, worse, begin not to pay attention to road conditions and hazards. There is no such thing as a semi-driverless car!
The aviation industry discovered this to its costs when pilots, accustomed to overreliance on autopilot systems, failed to react quickly or correctly in the face of unexpected situations. The Air France A 340 disaster in June 2009 over the Atlantic is the most deadly example of this phenomenon.
In effect, Level 3, whereby the car is in control most of the time but the driver has to stay alert the whole time, has been found to be unworkable. That is why Waymo etc gave up trying the gradual approach and are now attempting –without much success but that is another story – to go directly to full driverless capability.