Tesla recently announced a milestone in the race to deliver the world’s first autonomous vehicles, but are we really ready for the age of the self-driving car?
In late October, Tesla began rolling out beta testing for its Full Self-Driving‘ (FSD) software to a small number of customers. However, the arrival of the technology has been met with widespread concern from onlookers who don’t feel that the world, and artificial intelligence and machine learning, is ready to begin driving solo on the world’s roads.
(Image: Grand View Research)
With the self-driving vehicle market set to increase at a CAGR of 63.1% over the next decade, it’s clear that Tesla is looking to position itself at the forefront of innovation in autonomy ahead of its competitors in the likes of Uber and Google. But is Elon Musk attempting to send AI technology out into the big wide world before it’s entirely ready?
Technology That Could Do The Wrong Thing’
Commenting on the arrival of Tesla’s seminal Full Self-Driving software, The Verge acknowledged that Elon Musk has taken a bold move in opening his company’s autonomous vehicle technology to the public, but the tech website also noted that the beta looked scary as hell.
Customers who are brave enough to embrace the technology have been critical of the level of autonomy they’ve been given in trialling the advanced AI, however, with Tesla stating that drivers will need to remain vigilant at all times and be ready to take over control of their vehicle instantly, or else the car could do the wrong thing.
Speaking to NBC, Tesla Model 3 owner, Bill McGuire said that he finds FSD Terrifying, and that the branding is a lie, or at least a complete contradiction because of the level of human supervision that the cars require while driving autonomously.
Here, it’s vital to caveat the news of scepticism towards Tesla’s rollout with the fact that beta testing will inevitably involve a heavy amount of surveillance and vigilance, and the coming weeks and months will be fascinating to see how the technology takes on the test of real-world implementation. But according to a range of experts, the arrival of FSD technology is simply happening too soon.
The Gulf in Intelligence
The testing of self-driving cars in the past has resulted in accidents and even fatal collisions in the case of Elaine Herzberg, who was killed by a malfunctioning autonomous Uber car.
One key problem is that autonomous vehicles simply won’t be smart enough to come with all of the extensive variables that come with driving a car. In an article published in The Atlantic, Alexis Madrigal highlights 2017 MIT essays that claim there’s no short-term viability for self-driving vehicles because of the sheer number of edge-cases’ that an AI system will need to comprehend and respond to.
Factors like adverse weather conditions, low lighting, traffic diversions, various road markings and signage on an international scale and fast emerging obstacles can all challenge artificial intelligence in a way that it might be decades away from comprehending in a competent manner.
According to the Associated Press, there’s also the obstacle of autonomous vehicles interacting with human drivers. With people behind the wheel, there will be plenty of instances of human error that artificial intelligence will need to instantly account for and react to. For instance, during one recent test run, an Argo backup driver needed to override an autonomous vehicle after it stopped during a right turn at an intersection after the AI failed to figure out whether to go around a double-parked delivery truck.
Although Tesla has warned customers to remain alert to such confounding factors, there are few indications that AI will be close to displaying the adaptability to make quick calculations within these edge-cases any time soon.
Winning Over Customers
(Image: Statista)
Another problem that could affect the rollout of more AI-powered driverless vehicles is the fact that many drivers are simply unwilling to accommodate the technology. Some critics of autonomous vehicles have labelled their journeys as the most boring car ride of their lives, while some 73% of American drivers polled by the AAA have stated that they’re too fearful of stepping into a self-driving car.
The cost of embracing technology that will be both advanced and secure enough to think as intelligently as humans while driving as well as keeping users safe from cyber attacks and hackers may ultimately be another factor that keeps vehicles on roads altogether more manual for the foreseeable future. With this decade seeming set to be punctuated by a recovery from a global recession of an unprecedented magnitude, we’re likely to see more consumers opting for affordable used cars, and looking into car financing more to find a vehicle with technology-driven driving aids at affordable monthly rates.
This isn’t to say that we can’t get excited about a bright future for autonomous vehicles. However, at this stage, it seems fanciful to envisage the implementation of anything more complex than self-driving delivery vehicles for consumer goods, or some isolated taxi services by the end of the decade.
The road towards fully autonomous vehicles will certainly be a long one, but Tesla has at least shown the world that technology is taking the first pivotal steps towards making Full Self-Driving cars a reality. Sadly, it may be some time before you’ll be capable of taking a nap in your car while it whisks you to work in the morning, but anything’s possible in a future with Elon Musk at the wheel.
