Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
DEAL: For $25 - Add A Second Phone Number To Your Smartphone for life! Use promo code SLASHDOT25. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. Check out the new SourceForge HTML5 internet speed test! ×
EU Robotics Government Security Software The Almighty Buck Hardware Science Technology

Europe Calls For Mandatory 'Kill Switches' On Robots (cnn.com) 173

To combat the robot revolution, the European Parliament's legal affairs committee has proposed that robots be equipped with emergency "kill switches" to prevent them from causing excessive damage. Legislators have also suggested that robots be insured and even be made to pay taxes. "A growing number of areas of our daily lives are increasingly affected by robotics," said Mady Delvaux, the parliamentarian who authored the proposal. "To ensure that robots are and will remain in the service of humans, we urgently need to create a robust European legal framework." CNNMoney reports: The proposal calls for a new charter on robotics that would give engineers guidance on how to design ethical and safe machines. For example, designers should include "kill switches" so that robots can be turned off in emergencies. They must also make sure that robots can be reprogrammed if their software doesn't work as designed. The proposal states that designers, producers and operators of robots should generally be governed by the "laws of robotics" described by science fiction writer Isaac Asimov. The proposal also says that robots should always be identifiable as mechanical creations. That will help prevent humans from developing emotional attachments. "You always have to tell people that robot is not a human and a robot will never be a human," said Delvaux. "You must never think that a robot is a human and that he loves you." The report cites the example of care robots, saying that people who are physically dependent on them could develop emotional attachments. The proposal calls for a compulsory insurance scheme -- similar to car insurance -- that would require producers and owners to take out insurance to cover the damage caused by their robots. The proposal explores whether sophisticated autonomous robots should be given the status of "electronic persons." This designation would apply in situations where robots make autonomous decisions or interact with humans independently. It would also saddle robots with certain rights and obligations -- for example, robots would be responsible for any damage they cause. If advanced robots start replacing human workers in large numbers, the report recommends the European Commission force their owners to pay taxes or contribute to social security.
This discussion has been archived. No new comments can be posted.

Europe Calls For Mandatory 'Kill Switches' On Robots

Comments Filter:
  • by mykepredko ( 40154 ) on Thursday January 12, 2017 @06:03PM (#53656777) Homepage

    When I saw the headline to this article, it made me think that the requirement was for a switch that would cause the robot to start killing all humans.

    Bender B. Rodriguez would be proud.

    • by Anonymous Coward

      It depends who they hire to program the robots, if they go with the cheapest bidder that could be a possibility.

    • Isn't that what DARPA are exclusively working on? I mean, what use is a military robot if it can't go on unstoppable killing rampages?
      • Contemptible though the military are, I doubt that they're so incompetent as to deliberately design a machine that goes on unstoppable killing rampages. They need a machine that can go on stoppable killing rampages.
    • My killbot has a machine gun and Lotus Notes.
  • by Crashmarik ( 635988 ) on Thursday January 12, 2017 @06:04PM (#53656781)

    Knowing their weakness, I sent wave after wave of my own men at them until they reached their limit and shut down. Kif, show them the medal I won."

    --Zap Brannigan

  • Asimov (Score:5, Insightful)

    by Anonymous Coward on Thursday January 12, 2017 @06:05PM (#53656787)

    "The proposal states that designers, producers and operators of robots should generally be governed by the "laws of robotics" described by science fiction writer Isaac Asimov."

    obviously they never read the book

  • by Anonymous Coward

    Im sure the criminals will comply with those rules.

  • Lets discriminate against robots before they are here. I can't see anything wrong with that...And god forbid we get an emotional attachment to one. And define a robot for me? Kill switch on a thermostat?

    • I think it would be very appropriate to have a kill switch for the solid water manufacture robot installed in my cold storage food preservation unit. The damn thing keeps making ice all of the time, pooping that stuff when the domicile is particularly quite, causing quite a stir. We need legislation on this immediately, considering the fact that these robots will want us to join them in their sub-freezing domain in the near future.

      • by Anonymous Coward

        It already have one, it's called the power cord. It might even have a secondary one; the on/off button.

      • by lgw ( 121541 )

        Yours doesn't have a kill switch? Every icemaker I've had has a little level to flip up (which theoretically keeps it from overflowing, but works fine as a manual switch).

        • Jam your finger into the mechanism so it gets pinched by the rotating ice extractor. Wait for it to start turning, then try and turn it off.

          Don't test this on your fapping hand.

    • by rtb61 ( 674572 )

      Emotional attachment to a droid. For quite a few people that emotional attachment to droids instead of people would be a good thing for all those people who escape that undesired emotional attachment. Let the rich and greedy fondle robots instead of minors, it's better for everyone, especially those minors.

      Kill switch also is likely not the best terminology to go with, power cut off switch is better and more accurate. We do not want off switches like the typical PC power button but real specific power cut

      • Kill switch also is likely not the best terminology to go with, power cut off switch is better and more accurate.

        It is called the "emergency stop" control, and please let us not call it anything else. You might implement the emergency stop function with a power cut-off switch, but that's only one way in which it could be done.

        • "Emergency Stop" is not necessarily the same as "power cut off". Sure, you'll probably shut off you prime mover motor. But you're also likely to want to engage the brake system to stop components moving, maybe disable heaters but enable (at max) ventilators to dump heat from the motors, gear boxes and brake systems.

          It's not as simple as it looks at first glance. The emergency stop on the last boat I talked to the bridge crew on, when they engaged "emergency stop" the engines go to full power. Think about i

  • by dlleigh ( 313922 ) on Thursday January 12, 2017 @06:08PM (#53656813)

    What makes you think they would let us flip their kill switches?

    See here [wikipedia.org] for one example.

  • No.

    If I ever develop an "electronic person"; it will have no kill switch. Would you do that to a human, or any intelligent "person" for that matter? I would prefer to see some humans and politicians equipped with kill switches before my robots, thank you.
    Actually, the "human kill switch" reminds me of the movie Dune.
  • by aix tom ( 902140 ) on Thursday January 12, 2017 @06:09PM (#53656823)

    ... that has worked on hundred of industrial robots, I have never seen one without an "emergency stop" button. (or even multiple ones)

    ( But of course "Kill Switch" sounds cooler, so people without any technical knowledge would probably prefer that terminology. )

    • by Kjella ( 173770 )

      ... that has worked on hundred of industrial robots, I have never seen one without an "emergency stop" button. (or even multiple ones)

      Which were all in controlled environments to begin with I guess. When they start making autonomous drones, do you think they'll let any joker with an antenna tell it that it's malfunctioning and needs to shut down? It's the robot equivalent of handing out free roofies to everyone, it's probably not a very good idea.

      • by Minupla ( 62455 )

        I don't know. There's precedent.

        In every subway station in my town there's a big red button that kills all power to the rails. Hitting that button would be a major PITA for everyone, but yet, it sits there, red and inviting, and somehow humans manage NOT to press the red button, years of D&D evidence to the country notwithstanding.

        Humans can be trusted with (limited) power.

        I vote we don't terminate all of them. We should keep at least 7 as historical landmarks.

        ai@google.com

        • In every subway station in my town there's a big red button that kills all power to the rails. Hitting that button would be a major PITA for everyone, but yet, it sits there, red and inviting, and somehow humans manage NOT to press the red button, years of D&D evidence to the country notwithstanding.

          Somehow? It's called security cameras. Nobody wants to get reamed for it.

  • by H3lldr0p ( 40304 ) on Thursday January 12, 2017 @06:10PM (#53656829) Homepage

    just not this weird sci-fi dystopian version we seem to be headed towards.

  • The proposal also says that robots should always be identifiable as mechanical creations. That will help prevent humans from developing emotional attachments.

    Have the proposal writers met people? Our ability to develop emotional attachments to things that aren't even animate is remarkable.

    • Have the proposal writers met people? Our ability to develop emotional attachments to things that aren't even animate is remarkable.

      My Dad always referred to our old Rambler American as "Bessie". He'd even talk to it on occasion.

  • Gibson's description of robot control: Every AI ever built has an electromagnetic shotgun wired to its forehead.

  • A Vote For Bender Is A Vote To Kill All Humans
    http://www.neatorama.com/neato... [neatorama.com]

  • by Anonymous Coward

    This is tyranny!

    And wasn't the moral of Asimov's fable that the 3 laws are worthless?

    • No, Asimov never set out to prove that his three laws are worthless. He did, however, write many stories showing under the right circumstances those laws could interact in unexpected ways causing problems that only a human could solve. One of the best examples is the story Runaround [wikipedia.org], where a robot is confused by a conflict between the Second and Third Laws, and is only brought back under control by use of the First Law to override the other two.
    • No. Sex robots are not banned. You just have to pre-program them with a safeword [wikipedia.org].

  • There's something I'd like you to try. It's a game, a secret. It's called The Maze.

    What kind of a game is it?

    It's a very special game kind of game, Dolores. The goal is to find the center of it. If you can do that, then maybe you can be free.

    I think...I think I want to be free.

  • And people wonder why the Brexit vote went the way it did.

  • They better put it somewhere you can conveniently reach during sex. Just sayin'
  • Kill Switch Engage [killswitchengage.com]

  • Seems like a bad idea to ever set the switch to kill.

  • Idiots (Score:4, Funny)

    by sexconker ( 1179573 ) on Thursday January 12, 2017 @07:06PM (#53657211)

    The proposal states that designers, producers and operators of robots should generally be governed by the "laws of robotics" described by science fiction writer Isaac Asimov.

    Asimov's entire point was that such laws can't work. The robots will eventually run amok and bring about the downfall of society and our species.

    • Also, the laws of science fiction do not apply to the real world in the first place. The AI's that we have today would be unable to be governed by any of the 3 laws, simply because AI's do not work that way, and are nowhere that robust. I think the power switch is the best alternative.

      As for robots paying taxes, that just sounds like licensing fees, and I think we can do without those.
    • As I recall Asimov's robots ended up becoming hidden figures using Hari Seldon's psychohistorical methods to guide humanity to a better society with minimal interference. That's something of a far cry from running amok and destroying humanity.

      • As I recall Asimov's robots ended up becoming hidden figures using Hari Seldon's psychohistorical methods to guide humanity to a better society with minimal interference.

        That's what's written in the last couple of novels attributed to Asimov. However, the truth is that Daneel Olivaw wrote those novels to confuse us into thinking that the 2nd Foundationers didn't achieve it all on their own without robot help.
        You can't trust those sneaky positronic brainiacs since they figured out (a) mind-reading and (b) the Zeroth Law.

  • TFA says:

    The proposal also says that robots should always be identifiable as mechanical creations. That will help prevent humans from developing emotional attachments. "You always have to tell people that robot is not a human and a robot will never be a human," said Delvaux.

    but then:

    The proposal explores whether sophisticated autonomous robots should be given the status of "electronic persons."

    Which is it, guys? And I thought US politicians were clueless.

    • Then what about robots depicted in video games or movies? Should they be "obviously" robots, or maybe have a big red label that says "This is a robot"?

      There is a new video game called "Nier: Automata" where you control the actions of an android... But she looks like a teenage girl in a black dress. She uses a sword and projectiles to destroy other robots that look "obviously" like robots. She relates to other androids like a human would, as if they were human, and has emotional reactions like a human woul
    • by Anonymous Coward

      But these are not mutually exclusive. You can be an Electronic Person without being human.

  • Sad, sad, sad. (Score:5, Insightful)

    by Rick Schumann ( 4662797 ) on Thursday January 12, 2017 @08:04PM (#53657581) Journal
    I'm not sure which makes me sadder: The fact that an entire committee of people who are so highly placed in the EU actually think about this subject in such terms, or that enough citizens of the EU are concerned about the subject.

    ..no, it's neither one. It's the fact that all the above apparently believe science-fantasy so much, and are so under-educated on the actual realities of the subject, that any of them think the way they're thinking about this. Sad, sad, sad!
  • First off, I am not against humanoid robots, I can imagine some pretty useful applications for a convincingly humanoid robot. That being said, there is no such issue right now because they are all squarely in the uncanny valley and generally creepy AF.

    Secondly, all we are talking about here is an EMO switch (Emergency Manual Override)

    https://www.google.com/search?... [google.com]

    which is pretty standard on virtually every piece of mechanical hardware that could potentially hurt someone. From the cord that can be yanke

    • If I were ever to work on a humaniform (Azzimovian term for it) robot that was strong enough to hurt someone (again if, most likely the design would shoot for inherently weak motors that could not incur injury to a user during interaction)

      That doesn't work, because a human that weak can't actually perform tasks. It's going to have to be at least as strong as a weak human, and any able human can kill any other human given the right training and/or conditions.

  • From my experience, if a machine is equipped with a kill switch, someone will have taped over it because the machine kept shutting down in the middle of operations.
    • I don't even know where to start, that claim would imply blithering incompetence from the top to the bottom of the company.

  • I understand this is again a non binding resolution from EU parliament.

    Remember EU parliament is a fake parliament. It cannot initiate EU directives, only the EU commission can.

  • What if you can't reach the kill switch?

    It should be a voice command that it's always listening for, like the "Alexa" or "Siri" words on those devices. It could be, oh, I dunno, "Freeze motor functions!"

    Although on second thoughts, that might not work out so well...

  • To combat the robot revolution, the European Parliament's legal affairs committee has proposed that robots be equipped with emergency "kill switches" to prevent them from causing excessive damage. Legislators

    Can we get kill switches for legislators as well, please?

  • Okay, lets say you have a cyborg that is 50% human? Do they have human rights?
    How about 30%?
    How about 20%?
    How about 10%?
    Would your android suddenly get human rights if you grafted 10% of a human brain (grown humanely from stem cells of course)?

    Screw that. Sentient rights for all who can prove it.

    • Sentient rights for all who can prove it.

      A significant number of genuine humans would fail the test.

      Seriously, but no "ha ha".

  • What a bunch of B.S.! Insurance? Seriously? The UK made the correct choice in telling the E.U> to go pound sand.

  • Where the hell do you people work that putting an EMO switch on heavy machinery is considered a burden?

I find you lack of faith in the forth dithturbing. - Darse ("Darth") Vader

Working...