I’m a cyclist. Will the arrival of robotaxis make my journeys safer?

After having placed their profession in several American and Chinese cities for years, driverless taxis are on the way to London.

As a cyclist, Londonian and journalist who has spent years covering the AI ​​falls, I am a little nervous. However, given the frequency to which I was struck by unexpected human drivers in London, part of me is carefully optimistic.

At the end of the day, it comes down to this: will I be better surrounded by tired, distracted and angry humans, or unpredictable and imperfect AI?

The British government has decided to allow companies like Uber to direct autonomous service pilots “in taxi and bus” in 2026. Then, in the second half of 2027, things will increase quickly as the Automated Vehicle Act becomes the law, giving industry an appropriate legal framework. Above all, this law is due to car manufacturers responsible for accidents rather than occupants.

The government claims that driver -free cars could in fact improve security, since human error contributes to 88% of all road accidents. And there are many: there were 130 dead on London motorways last year, which includes 66 pedestrians and 10 cyclists. Globally, around 1.2 million people die each year on the roads.

I bike everywhere in London and it gives you an overview of the problem. I saw drivers read, eat cereal bowls and watch movies. I was crushed from behind to Red Lights at least four times. They say that a thing that is missing in AI is creativity, and in the field of bad behavior, humans really have a flair.

Meanwhile, AI is not attempted with a text, does not take drugs or drink, nor soda. He does not make turns without checking the dead angles, because there is nothing like a blind spot for a machine that has dozens of sensors.

Yes, there are very disturbing examples of driverless cars that do not stop for pedestrians and kill them. This is rightly news. But we have become so Inu -réés on the road deaths that the most numerous reports involving human drivers are barely – more than four people a day die on British roads, on average.

The Robocar safety problem is delicate. In my opinion, not a single death on the road should be tolerated, but from a pragmatic point of view, if the AI ​​can drive the same number of kilometers and kill fewer people, then there is a strong argument that we should not perfect the enemy of progress.

Indeed, studies have shown that driver -free cars tend to be safer than those piloted by humans, although this record lowers low lighting and when making turns – barely a rarity in city driving – and there are concerns about the under -declaration of accidents.

We are also counting on technological companies to make Robotaxie safe, and there could be a conflict of interest between profit and the greatest good. We have already seen morally repugnant efforts to transmit the problem to pedestrians by suggesting that they wear electronic sensors to spread their presence to such machines.

Regarding cyclists, do technological companies ensure that they receive 1.5 meters of space when their Robocars exceed them, or do they decide that as long as a cyclist is not overturned, it’s ok? The latter would improve travel times by car in a lively city and terrify and cyclists in danger. To what extent will Robocars get out of the secondary roads to reach the loaded routes? Do they wait for pedestrians to completely cross the roads or continue to drive at a rate that encourages the sad little trot that we do to appease the drivers? These are all parameters that can be changed, and there will be a commercial dispute between safety and speed of travel.

Even if companies act altruistic, AI is not deterministic. We cannot perfectly predict how he will behave in a given situation. Just as you cannot guarantee that a chatbot does not listen to the glue as an ingredient in a pizza recipe, you cannot be completely sure that an autonomous car will not decide that a pedestrian is a shadow and made them pass. It is not practical, or reassuring to hear, but it is a fact.

In truth, I do not trust the AI ​​to operate a car near me. I also don’t trust people to do it. But while humans are about as good as possible, AI has the potential to improve quickly. Autonomous taxis tests in London will provide valuable training data, with the potential to improve safety in a virtuous circle. In the long term, if I had to choose, I opt for AI drivers.

That said, the harsh reality is that a few tonnes of steel, five armchairs and 100 computers on four wheels will never be a reasonable, completely safe or efficient means of urban transport. Such taxis are such a mediocre solution to transport to the London of tomorrow as those of man today.

Electric bikes and safe cycle paths are greener and can ensure that people from A to B faster, while buses can transport 80 people in the space absorbed by two SUVs. But there is no beneficiary margin for this for Big Tech, right?

Matt Week

What I read

How does music work Speaking of the leader of chefs David Byrne.

What I look at

Horror movie Bring it back (through your hands tight on the eyes at points, certainly).

What do I work

Prepare a lot of cuttings in the garden to fill bare places next spring.

Matt Sparkes is a technological journalist at New Scientist