<center>https://fsmedia.imgix.net/a6/7a/52/85/9ea9/441d/88f1/8031fc0a6aea/pepper-the-robot-attends-the-2017-new-yorker-techfest-at-cedar-lake-on-october-6-2017-in-new-york-c.jpeg?rect=0%2C188%2C3600%2C1797&auto=format%2Ccompress&dpr=2&w=650</center>
## Could you think of any circumstance where it would be a good idea for a robot <b>not to follow orders?</b>
<center> </center> Obviously there are times where <b>blindly following what anyone tells you to</b> without considering the implications for your own well-being, might not always be the best approach to take. 

<center>https://www.iso.org/files/live/sites/isoorg/files/news/News_archive/Ref2169/Ref2169_romeo.jpg/thumbnails/800</center>
What if someone with *occasional memory loss* for example, owned a robot that provided self-care services in the future and they asked that robot to perform a task <b>that they had already completed</b> that day, you might want that robot to instead refuse wasting their time being obedient to that request. Or what if a robot is asked to take a purse from a lady who is seen sleeping on a bench; it wouldn't be wise for them to be programmed to follow *any order that is given* because sometimes that could bring harm to themselves or others, or might bring undesired consequences.

## These robots are sophisticated tools that are being programmed for a variety of useful purposes but it's important to try and prevent unnecessary  harm and that might include enabling those robots to say no on occasion. 
<center></center> If they are being asked to perform a task that might cause harm, be impossible, or to do something that is unethical etc, it's important they can exercise caution and this is why researchers have been busy trying to teach these machines how to exercise a choice.

<center>https://assets3.thrillist.com/v1/image/2414518/size/sk-2017_04_article_main_mobile.jpg</center>
## In one experiment (see video below), researchers have been working on developing <b>robotic controls that make simple inferences</b> *[according to human commands](https://theconversation.com/why-robots-need-to-be-able-to-say-no-55799)* that are given. 

<center></center> In the video below you can see researchers asking the robot to walk off of a table, but the robot refuses insisting that there isn't any sufficient support and that it will fall off, but the human giving the orders reassures the robot that he will catch him so the [robot then responds by walking off the table](https://spectrum.ieee.org/automaton/robotics/artificial-intelligence/researchers-teaching-robots-how-to-best-reject-orders-from-humans).

<center>https://youtu.be/0tu4H1g3CtE</center>
# Researchers have also been interested in *[how people might respond](https://hrilab.tufts.edu/publications/briggsscheutz14ijsr.pdf)* when a robot rejects a command that is given. 
<center> </center> To investigate their potential response, researchers set up an experiment that involved adult test subjects asking robots to knock over towers of aluminum cans that they had made. 

When the participants entered the room, the robot had finished building their can tower and verbally expressed pride with the accomplishment to the test subject, about how happy they were to finish and that it had taken them some time to complete.

With [one group of participants researchers](https://hrilab.tufts.edu/publications/briggsscheutz14ijsr.pdf) noticed that the robots complied with every order to *knock it down*. But with another group of participants, they got much different results. When asked a first time to knock down the tower, the robot would reply <b>it reiterated that *it had just built the tower*</b>, when asked again a 2nd time to comply, it <b>argued that it worked really hard on completing the tower</b>, and when asked a third time to comply, the robot knelt down and started making sobbing noises, crying no in response. On the 4th issue of the order to knock it down is when the robot finally slowly walked over and complied. 

Researchers found that *a number of participants who had observed the protests from the robot* attempting not to comply with the command, had left the tower standing. This supports the notion that a robot might be able to discourage people *from following a particular course of action*. 

Pics:
[Pic1](https://www.inverse.com/article/38669-robots-to-buy-for-future)
[Pic2](https://www.iso.org/news/Ref2169.htm)
[Pic3](https://www.seeker.com/socially-assistive-robots-could-make-you-healthier-not-jobless-2315669739.html)

<center>https://steemitimages.com/DQmaW48YhsgLS2fxaYkXCuon3tFpQEEh1tSUCU3jjJXv7RT/%40doitvoluntarily.gif</center>

## [Getting Robots To Pack The Groceries](https://steemit.com/technology/@doitvoluntarily/getting-robots-to-pack-the-groceries)