Industry News: Waymo Scraps Plans For Human Intervention In Dangerous Situations

[William Maley]

A number of companies that are working on autonomous driving systems are implementing features that require a driver to take control in hairy situations. But Waymo, the self-driving unit of Alphabet (Google's parent company) recently revealed they have stopped developing these features. The reason is simple, drivers lacked the awareness to take over driving.

Waymo's CEO John Krafcik told reporters yesterday the decision to scrap this technology comes down to experiments done back in 2013 with Google employees that were filmed. The videos shown for the first time outside of Google showed people messing on their phones, putting on makeup, and even someone deciding to take a nap.

“What we found was pretty scary. It’s hard to take over because they have lost contextual awareness,” said Krafcik.

After the napping incident, plans for a driver to intervene were scrapped. Since then, Waymo has been working on technologies that didn't require human interaction. As such, Waymo's Chrysler Pacificas running in Phoenix, AZ only have two 'drive' buttons: one to start a drive and the other to tell the vehicle to pull over at its earliest convenience.

Source: Reuters

View full article

[David]

We'll this is a no brainer, how can one expect a person to take over control of an auto when they have not been paying attention the whole time to what is going on around them. Common sense or a lack of one at Waymo. Crazy Alphabet / Google Peeps.

[riviera74]

4 hours ago, dfelt said:

We'll this is a no brainer, how can one expect a person to take over control of an auto when they have not been paying attention the whole time to what is going on around them. Common sense or a lack of one at Waymo. Crazy Alphabet / Google Peeps.

They were not crazy.  They simply followed the data and observations.  Besides this happened four years ago.  I am sure that human intervention was never going to be implemented once self-driving took over.

[Drew Dowdell]

Computers are smarter and faster than humans.

[Robert Hall]

16 minutes ago, Drew Dowdell said:

Computers are smarter and faster than humans.

Problem is they are programmed by humans and always have faults..

[Drew Dowdell]

Just now, Cubical-aka-Moltar said:

Problem is they are programmed by humans and always have faults..

Yet fewer faults than humans. 

[FAPTurbo]

On 10/31/2017 at 8:46 AM, dfelt said:

We'll this is a no brainer, how can one expect a person to take over control of an auto when they have not been paying attention the whole time to what is going on around them. Common sense or a lack of one at Waymo. Crazy Alphabet / Google Peeps.

this was common sense... Waymo's engineers very likely knew how it'd turn out this way, but then they wouldn't have the data needed to counter concern-trolling luddites who'd insist that the machines are out to kill us

[balthazar]

...lost contextual awareness.

The (theoretical) transition between driver-controlled & true AD cars is going to be an entertaining/madly dangerous time once those who become 'programmed' to nap get back behind the wheel of a non-AD car. That's going to be, like, a 20 year window, too.

Edited November 6, 2017 by balthazar

[oldshurst442]

2 hours ago, FAPTurbo said:

this was common sense... Waymo's engineers very likely knew how it'd turn out this way, but then they wouldn't have the data needed to counter concern-trolling luddites who'd insist that the machines are out to kill us

You sure you wanna open that Pandora's box, tho?

Greek mythology...

yes...

myths...

Not to be taken so bloody literal...there is a message to be learned in these Greek myths...

Read the lyrics in this song...

released in 1969...

all that in the song is happening now...the songwriter did not have to wait THAT long for all THAT to happen...just a mere 45 years...

 

Just a thought...

Edited November 6, 2017 by oldshurst442

[Robert Hall]

I wonder if the future is going to be a mix of Blade Runner (robots built to do the crappy work, no flying cars though), The Terminator (computers taking over and killing humans), and post-apocalyptic stuff like The Road....but seriously, being in software, I see how buggy and insecure software systems are on a daily basis, I certainly would't trust them to control the complete driving of cars, heavy trucks, etc any time soon...

I recently read After On, a fun read about Silicon Valley startup companies, social networking, AI and terrorism... 

[David]

14 hours ago, Cubical-aka-Moltar said:

I wonder if the future is going to be a mix of Blade Runner (robots built to do the crappy work, no flying cars though), The Terminator (computers taking over and killing humans), and post-apocalyptic stuff like The Road....but seriously, being in software, I see how buggy and insecure software systems are on a daily basis, I certainly would't trust them to control the complete driving of cars, heavy trucks, etc any time soon...

I recently read After On, a fun read about Silicon Valley startup companies, social networking, AI and terrorism... 

Too late, Saudia Arabia already gave citizenship to the first AI human.

http://www.businessinsider.com/meet-the-first-robot-citizen-sophia-animatronic-humanoid-2017-10

Interesting reading and will scare many who fear AI.

[Drew Dowdell]

19 hours ago, balthazar said:

 

 

The (theoretical) transition between driver-controlled & true AD cars is going to be an entertaining/madly dangerous time once those who become 'programmed' to nap get back behind the wheel of a non-AD car. That's going to be, like, a 20 year window, too.

It's scary when Albert goes from the Buick to the Honda and backs up without the rear-view camera and no parking sensors......