A Robot Recognizes Itself
Written by Harry Fairhead   
Saturday, 03 December 2011

A cute robot called QBO seems to have the ability to recognize itself. Does this demonstrate self-awareness? Or is it just a gimmick?

 

There are many ways of testing for the presence of intelligence and the best known is the Turing Test - in many ways the weakest test of all. When it comes to biological systems we have certain expectations generated by the fact that we are one. We are also reasonably confident that we have a graduated scale of complexity from the simplest animals processing the simplest neural systems to the most complex - that is presumably ourselves again.

What all this means is that we are allowed to do a little more than perform black box testing on electronic intelligences. We can demand not only a pleasing chat conducted via a keyboard and screen or a voice recognizer and speech synthesizer. We can demand a certain amount of communality of experience and state.

 

QBO

 

One of the things we can do is recognize ourselves in a mirror. When I point at my reflection and say "that's me" it isn't just a piece of physics but a tiny revelation of the state of being self aware.

I know that there is a me to be recognized.

Robots are often cute and they can work on human empathy to place themselves into a position of a pet or, perhaps more worryingly, a small child. This can go to the point where we are fooled into thinking that they are a concious entity similar to, but perhaps lesser than, ourselves. We can be fooled into thinking that they are self aware.

Now we have a video "QBO and the Mirror" that places this claim to self awareness clearly to the public. A small robot, Qbo, was placed in front of a mirror and after some learning algorithms had run their course it utters

"Oh. This is me. Nice."

So we have proof that robots can be self aware.

Watch the video and see what you think.

 

 

Even if you are told that QBO used object and face recognition, are you any more impressed?

It probably is more impressive to the general public who would have no idea what the inner workings fo the robot are.

Back in the late 1960s, there were many attempts to teach chimpanzees and other apes sign languages of various sorts. You can still argue today about how successful they were. Chimps did learn to use signs in a grammatical way, i.e. they learned some syntax for putting signs together but who knows if the syntax was advanced enough to call the use of symbols a language.

However one chimpanzee, Sarah, was filmed standing in front of a mirror and signing "that's me". When I first saw it the hair stood up on the back of my neck - it was a thrilling first contact.

In this case any viewer is left in no doubt that this is self recognition with a strong suggestion of being self aware.

When it came to watching QBO, no such emotional response was triggered, more an appreciation of the algorithms employed.

Was the distinction just because we do not know the way that the chimp's, or any brain, works? That ignorance allows us to extrapolate in ways that you simply cannot with the crude algorithms in today's robots. However, there is a little more to it than this. The chimpanzee brain is in the same series as our own. It arguably isn't as good but it is the same stuff. For this reason we can empathize and we are allowed to read more into QBO's weak attempt at proclaiming self awareness than is justified.

 

QBO

 

Still, there is a lot of scope for misunderstanding and I'm sure that you will read other news stories that are more credulous in the near future.

More Information

Having searched we can't find the original BBC Horizon documentary of Sarah signing in front of a mirror. If you are interested in seeing some of the early work on animal languages, you might like to view another documentary in the same series:

 

  

 

 

To be informed about new articles on I Programmer, subscribe to the RSS feed, follow us on Google+, Twitter or Facebook or sign up for our weekly newsletter.

 

 

Banner


Google Updates Responsible AI Toolkit
01/11/2024

Google has announced updates to the Responsible Generative AI Toolkit to enable it to be used with any LLM model. The Responsible GenAI Toolkit provides resources to design, build, and evaluate open A [ ... ]



GitHub Universe AI Announcements - Copilot And Spark
30/10/2024

GitHub has announced several improvements for developers at Universe, its annual conference. Developers will get multi-model Copilot and GitHub Spark, an AI-native tool for building applications in na [ ... ]


More News

Last Updated ( Friday, 23 June 2017 )