AI update deep dive: AI in autonomous driving

The idea of ​​making cars drive autonomously is almost as old as the first cars. Author Clemens Gleich even speaks specifically about the fact that carriages were quite autonomous – namely the horses at the front. This deep dive on the AI ​​update is about how artificial intelligence makes autonomous driving possible and what similarities the hype around AI has to that of autonomous cars. Our heise cars expert Clemens Gleich knows that during the first driving tests there were fears that the systems would put an incredible number of people out of work. With truck drivers, taxi drivers and other drivers, this could have affected a very large group of people. But little has happened so far. And Clemens Gleich also has a little doubt as to whether a human might not be cheaper than an autonomous robotaxi like the ones already driving around San Francisco.

Advertisement

Many autonomous driving tests are underway in California. But there is also protest. For example, activists recently disabled autonomous cars by placing commercially available pylons on their hoods. That’s enough to confuse the systems to such an extent that the cars simply stop.

Cell phone networks are often overloaded at large events, as was the case at a music festival in San Francisco. What no one had sufficiently considered beforehand also disrupts the signals of the autonomous cars so much that they also stop. Despite the technology on board and plenty of AI, in cases of doubt the cars sometimes send messages home, where a human then makes the decisions quite simply.

We also talk to Clemens Gleich about who is actually liable and on what basis the AI ​​makes its decisions: “Many decisions are determined algorithms.” A zebra crossing means stop. The AI ​​does calculate probabilities, but too much leeway doesn’t make sense, explains the expert.

(emw)

To the home page
#update #deep #dive #autonomous #driving