Cars not yet 'smart' enough to be in control

On the one hand, self-driving cars are long overdue. How can we send unmanned spacecraft to roam the solar system and beyond and let the auto-pilots take care of much of the air travel but leave road traffic to the skill, or lack thereof, of human drivers? 

On the other hand, plenty of evidence has shown that we trust computers at our own peril.

Like it or not, the world is moving closer to the day when vehicles driven by humans become something that catches the eye. Smart cars are already here, and the only major obstacle preventing them from getting smarter to the point of full-scale self-navigation has to do more with human pride and morality than technicality. 

But there are those who don’t want to wait for the safety debate to end, like the manufacturers of self-driving trucks that are set to revolutionalise haulage companies in the United States.

Ironically, the safety concerns that deterred the evolution of self-driving cars influenced the innovation of the Future Truck 2025. According to the US-based National Highway Traffic Safety Administration, 333,000 trucks were involved in crashes throughout America in 2012, killing nearly 4,000 people, most of them travelling in passenger vehicles.

The Future Truck idea is simple enough. It’s equipped with a highway pilot automated system that will enable it to cruise on highways on its own. On busy urban roads, humans retake the steering wheel. 

According to someone involved with the development, the Future Truck “never gets tired. It’s always 100 per cent and sharp. It’s never angry. It’s never distracted. So, this is a much safer system”.

Safe to a certain extent, argues the conservative, pro-human camp. The sceptics believe that while the computers are better in certain conditions and in terms of reaction time, they are far from ready to match the human level of decision making and our ability to react in the abstract in unpredictable situations. When driving is concerned, there are certain “grey area” situations that can dumbfound the computers, they say.

As if humans are good at making the “grey area” decisions, the other camp shoots back. Accident statistics may be useless in this debate at the moment, because we only have records of human, not computer, errors.

A powerful yet probably politically incorrect argument against self-driving cars has to do with human selfishness. Suppose there are two options, the first will hurt or kill your travelling family members but the second may send a school bus off the cliff. What would you do and what would the computer do? 

Yet the pro-technology camp insists that not having to come to that painful crossroad is the whole point of having computers plot safety navigation from the get-go. 

The debate will go on, but it’s unlikely to resist the push of technology very much. Backup cameras, adaptive steering, lane-guidance systems and even automated parallel parking exist in today’s cars, which means they are already “smart”. Many car manufacturers are working on self-driving prototypes and the innovators involved have said the same thing – they are doing it for greater road 

safety.

All in all, we have seen nothing yet. Self-driving vehicle prototypes being tested nowadays are like the very first versions of cellular phones or flat-screen TVs, meaning there is plenty of room for improvement, and the improvements will come cascading down assembly lines soon.