Dumb and dumber
Intelligent cars, not-sointelligent humans. Should the latter be reprogrammed to suit the former?
The best accompaniment for robot cars is … robot people. As motor companies and their technology partners pump billions of dollars, euros, yen, yuan and pounds into self-driving vehicles, one obstacle stands obstinately in their way: the human being.
Artificial intelligence is no match for human intelligence — or, rather, the lack of it. The first recorded pedestrian death involving a fully autonomous vehicle happened in the US this year when a self-driving Uber car hit a woman in Tempe, Arizona. The incident added to growing public concern about the safety of such vehicles.
A recent US survey found that nearly 50% of consumers polled said they would never buy a fully autonomous vehicle. That’s up from 30% two years ago. The survey, by Cox Automotive, found public confidence has been shaken by a series of mishaps during testing of autonomous vehicles.
Another research firm, Gartner, has shifted autonomous driving technology into the “trough of disillusionment” category of its trends analysis.
Most respondents in the Cox survey said they preferred current automotive technology to the future kind. Only 16% would feel safe in a fully self-driving vehicle that did not allow them to take back control.
While 75% of the sample said fully-autonomous vehicles need real-world, in-town testing in order to be perfected, 54% said this should not take place anywhere near them.
Defenders of self-drive technology say there’s nothing wrong with the technology itself. The problem is unpredictable people: jaywalkers and motorists who don’t adhere strictly to traffic rules. The Arizona incident involved a woman walking her bicycle across a road at night. But a preliminary report says because she was outside the official pedestrian crossing zone, the Uber’s sensors failed to accurately interpret the data, causing what safety officials have quaintly called a “false positive”.
US publication Automotive News quotes a scientist as saying: “The problem with most of the computer vision systems that selfdriving cars use is that they simply put a boundary box around an object and apply a label — parked car, bicycle, person — without the ability to analyse anything happening inside that box.”
One prominent automotive technologist says the answer is for people, particularly pedestrians, to be retrained. This has caused a fierce backlash from some other scientists, who accuse their peer of arrogance for suggesting the world should bend to accommodate flawed technology.
This led Uber and Waymo, two of the companies leading the way in self-driving technology, to issue statements that their technology is being designed to handle the world as it is, not as they would like it to be.
Several companies are stepping back from earlier autonomousdriving production deadlines amid growing acceptance that these were unrealistic. Commercial driverless trucks were predicted to be ready for operation by 2020. Daimlertrucks now says it will be at least 2023.
Despite delays, spending on future technologies remains undimmed. Toyota is investing $500m in Uber in pursuit of autonomous ride-sharing. The Japanese vehicle manufacturer has previously ploughed $1bn into
Grab, an Asian rival of Uber.
Japan is already home to what it claims to be the world’s first autonomous taxi service on public roads. The central Tokyo service carries passengers along a fixed
5.3 km route between the central railway station and a popular entertainment district. For now, the pilot project — which has a driver in the vehicle in case of emergency — offers four daily return trips but the plan is for fully commercial operations by 2020, when Tokyo is due to host the Olympic Games.
Other commercial autonomous projects are also up and running. In the Californian city of San Jose, food retailers have partnered with a technology company to offer driverless grocery deliveries to nearby residents.
Japan also hopes to take the lead in flying cars. The government has announced plans to incorporate them in its public transport system. It is talking to companies like Airbus, Boeing, Uber and Japan Airlines.
Besides helping alleviate traffic stress in — and over Tokyo — the project would also provide transport to remote areas of the country and offer tourism services.
The country’s trade ministry has reportedly allocated $40m for the development of flying-car components. It said in a statement: “The Japanese government will provide appropriate support to help realise the concept of flying cars, such as the creation of acceptable rules.”
Move over, people: Are robots better suited to future driving?