No consciousness in control: Self-driving car crash, no one hurt

A car in self-driving mode (the one on its side in the photo) was just involved in a crash in Tempe, Arizona. Having recently driven 18,000+ miles in 3 months through every lower US state, I am skeptical of self-driving cars. Also, beating my chess computer at the highest setting yesterday morning made me trust self-driving artificial intelligence even less. Human drivers with consciousness have a superior ability to navigate unique situations.

I dealt with crazy stuff while driving across America. In one place the lines on a multi-lane road disappeared. The locals knew how many lanes there were, but I had no way to tell. Some traffic lights and signs were strangly oriented, oddly worded, or barely visible. Visibility dropped to a few feet on a freeway once due to massive rain. A rogue gust of wind almost pushed me off the road. There were deer, cattle, other animals and some terrible drivers.

Like a chess computer, an artificial driving intelligence may have a massive database of everything that has happened in the past, including the best reaction to each situation, and this may make it statistically a safer driver, but the real world is filled with stupid little exceptions. 

That’s the thing. AI driving is learning, but progress is rocky according to leaked documents. It may be easier to give a computer consciousness than to program reactions to the unlimited exceptions that  come up. So many variables easily handled by conscious humans come into play when faced, for example, with a split-second yield decision. 

Failure by another vehicle to yield is a very common situation. Would an alert human have avoided this accident entirely? 

Uber has suspended its fleet of self-driving cars while it investigates a crash in Arizona involving one of its vehicles. Photographs posted online show the car lying on its side next to another badly damaged vehicle. There were no serious injuries as a result of the incident.

The Uber Technologies car – a Volvo SUV – was carrying two engineers in the front and no backseat passengers … 
The incident is the latest blow for the car-hailing app which has been left reeling by a series of setbacks in recent weeks including the departure of its president, Jeff Jones. Initial police reports suggest the collision was caused by a person who failed to give way to the self-driving car rather than a malfunction by the Uber vehicle.
However, pending further investigation, the company has removed all its self-driving cars from the road in Arizona, as well as test sites in Pennsylvania and California – all three states where it operated the driverless vehicles

A spokesperson for the police in Tempe, Arizona, said the crash happened when another car “failed to yield” to an Uber car at a left turn. Josie Montenegro said: “There was a person behind the wheel. It is uncertain at this time if they were controlling the vehicle at the time of the collision.”

Uber self-driving cars always have a human in the driving seat who can take over the controls.

The latest incident follows recent difficult months for the firm, which has seen a number of high level executives quit and has faced criticism over workplace practices and ethics.
Meanwhile legal issues have been hampering progress with the testing of its autonomous vehicle technology in California.

Link

Self-driving cars aren’t ready yet. They may never be, until we have quantum computers with something like consciousness.

… there have been accidents, including a fatality in Florida in May when a truck struck a speeding Tesla that was on autopilot.

An investigation found no safety-related defects with the autopilot system, but concluded that the driver may have had time to avert the crash if he had been paying closer attention.

Link

Here’s a self-driving car caught running a red light

img 0574 300x229

“But even though Uber said it had suspended an employee riding in the Volvo, the self-driving car was, in fact, driving itself when it barreled through the red light, according to two Uber employees, who spoke on the condition of anonymity because they signed nondisclosure agreements with the company, and internal Uber documents viewed by The New York Times. All told, the mapping programs used by Uber’s cars failed to recognize six traffic lights in the San Francisco area. “In this case, the car went through a red light,” the documents said.”

Link

Some see consciousness, including imagination, as the evolutionary answer to environmental unpredictability. In other words, the reason we imagine things is to be as prepared as possible for the unpredictable. When will AI drivers be given imaginations? Careful what you wish for. It’s a double edge sword: too much imagination makes you detached from reality, that is, crazy.

Update: 

Uber’s self-driving cars are back on the road, following a weekend crash in Tempe, Ariz. that grounded test cars in Pittsburgh and Arizona.
Tempe police confirmed to CNNTech the self-driving Uber vehicle involved in the rollover accident Friday night was not at fault, and there were no life-threatening injuries.
The self-driving car tests resumed in both cities Monday, after a temporary halt to investigate the accident.
In addition, Uber’s self-driving car testing resumed in San Francisco, the company told CNNTech.
Uber had been forced to stop testing in that city in December, after California regulators revoked the vehicles’ registrations. The company had skirted regulations by not applying for permits, and after it ignored warnings from the DMV, regulators halted its program.
Uber then moved its testing to Arizona, which has no special regulations for autonomous vehicles.
A spokeswoman for the California DMV confirmed Uber obtained a permit to test its autonomous vehicles on March 8. That permit covers two vehicles and 48 drivers, and each additional car and driver will need to be approved individually. …

Link

SOURCE b gif host truestrange