Recent news of so-called “autonomous” vehicles in Reno and Las Vegas have us wondering how soon they might debut in a place like Elko – and how they might be received by motorists here in Nevada’s outback.
In downtown Las Vegas, a driverless electric shuttle started operating on public streets this month under a pilot program. It carried up to 12 passengers and a human attendant along a stretch of Fremont Street. The vehicle has no steering wheel or brakes, and was guided by GPS without the need for lane lines to navigate the route.
In Reno, a University of Nevada project is getting underway to collect data that will enable an electric bus to learn how to drive itself. The buses will still have drivers – at least for now.
So, just how reliable are these vehicles and how close will they get to using human-like judgment when it comes to keeping us safe? It’s an important question, considering that a man was killed last year when his driverless Tesla slammed into a semi when it was unable to distinguish the truck’s trailer from the clear, blue sky.
Tesla CEO Elon Musk downplayed the danger, but elsewhere he has described the science of Artificial Intelligence as “our biggest existential threat.”
The technology that is rapidly enabling driverless vehicles to come out of the workshop and onto public streets was largely developed under Department of Defense programs. DARPA’s Deep Learning (DL) program is a result of combining artificial neural networks with “Big Data.”
“Due in large part to the DL revolution, we are in the surprising position that the rollout of self-driving vehicles is now more limited by the speed of policy change than by its technical readiness,” states an unclassified document authored by the nonprofit MITRE Corporation.
Nevada is one of the places where policy changes are enabling the real-time testing of driverless vehicles. Scientists prefer the term “autonomous” but such vehicles really do not have any freedom of choice – they can only respond to sensory data based on pre-programmed conditions. On the other hand, humans – who are truly autonomous – can set their own goals and strategies.
The difference was highlighted this month in an Associated Press report that looked at how driverless vehicles might respond to a “moral dilemma”:
“Imagine you’re behind the wheel when your brakes fail. As you speed toward a crowded crosswalk, you’re confronted with an impossible choice: veer right and mow down a large group of elderly people, or veer left into a woman pushing a stroller. Now imagine you’re riding in the back of a self-driving car. How would it decide?”
To make such a choice requires a leap from Artificial Intelligence to what scientists are calling Artificial General Intelligence, meaning “a human-like ability to pursue long-term goals and exercise purposive behavior.”
According to the Federation of American Scientists, this field using multi-layered neural networks is experiencing “broad and unforeseen successes.” For example, this month the Department of Defense demonstrated a swarm of autonomous micro-drones that “demonstrated advanced swarm behaviors such as collective decision-making, adaptive formation flying, and self-healing,” reported FAS.
Still, the organization concluded that “Sentient machines, let alone a revolt of robots against their creators, are still somewhere far over the horizon, and may be permanently in the realm of fiction.”
Driverless technology still has many obstacles to overcome, according to the MITRE report, which explains that vehicle behavior has to be based on some sort of map:
“Then, if it can see lane markings, it can figure out what to do — unless there has been recent construction, in which case it must decide which are the old misleading lane markings and which are the new ones. If it is snowing, or there is a road with no lane markings, what does it do? … More complex is the interaction with other things, from other vehicles, to pedestrians, to potholes, to careless children, to windblown garbage cans.”
These are real-world anomalies that humans are used to encountering, although not always successfully. To find out if machines can do it better, we need demonstrations such as the ones under way in Nevada.
We wish these brave, new inhabitants luck as they begin to populate our streets and highways. Just don’t expect us to wave as they drive by.