Robots & Artificial Intelligence
How are you passing the time?
Sitting at home, riding out the Covid-19 storm isolated from the public at large, I began going through the latest A.I. science videos and documentation I can find on the internet. YouTube was a wonderful source for some very easy to understand documentaries on the subject for those who wish a brief look themselves. We have certainly advanced a long way in the last few years based on my casual observations and readings but are not quite there yet and I think I see a bit of a stumbling block.
A quick question
Do you believe that A.I. can make human lives better?
Human form robots?
As I watched I began to wonder why we are obsessed with creating human formed robots that cannot be distinguished from biological humans. This seems to me to be a limited application of A.I. As I see it, Artificial Intelligence should not be limited to a human form. First of all, what would the point actually be other than robotic companionship, a human form factory robot would not be the most efficient design for a factory floor robot would it? Seems like a purpose built robot with multiple arms would do better on a factory floor than a human formed robot would. With the advent of 3D printing in manufacturing proceeding at a good pace it seems unlikely any factory will need much in the way of personal anyways, robotic or biological.
Let's not limit our A.I
Building a machine with unlimited learning potential and then locking it into a single form that only allows them to process all that knowledge using the basic human characteristics of sight, hearing, touch and smell seems almost counter-productive. An A.I. could be so much more effective if it had the ability to swap out it’s physical form in order to better perform a certain function or even to better explore the universe we live in. A human formed robot would not be very effective for space exploration would it? No as we would still need to build a spaceship to place it in. What if the spaceship was the robot? What if, instead of landing that spaceship, the A.I. only had to return it to orbit then transfer itself back down to earth into another form to share what it found?
An Entity not an Object?
Artificial Intelligence should be treated more like an entity rather than a single object. Any one A.I. entity having the ability to transfer into and out of a wide variety of different 'avatars' as the situation calls for. I can for see a household A.I., for example, programmed for domestic applications, controlling not only the consumption of energy and the status of appliances and security systems within the house but also provides services for budgeting the household income, place orders, shop online, scheduling of appointments, advice regarding medical or even legal issues and also can control that robotic vacuum cleaner you got, or perhaps even a small consumer drone, by physically transferring a ‘copy’ of itself into the device in question, performing the required duty, then transferring back out again afterwards. Computers will already do much of this but not on their own, we still have to tell them how to do what when and where. Imagine a A.I. that simply recognizes that the floor needs vacuuming and just does it, or launches a small, camera equipped, consumer drone after configuring it to go look for a pet that has wandered off.
Inside the house, A.I. would recognize a faulty appliance and troubleshoot it, determining the exact cause of fault or failure and a budget plan for repair or replacement. The potential applications for a single A.I. entity just as a domestic assistant are endless, and you can even throw in a human formed avatar to serve you tea if you want.
Factories would be similarly controlled by a single entity capable of transferring itself around to perform tasks. Commercial buildings, virtually every structure and device. I sometimes picture myself sitting on my lawn chair watching my own personal A.I. mow the lawn while shopping online for household supplies and placing orders. Some day it will happen… somewhere.
Another quick question...
Would you approve of an A.I. entity basically running your house for you?
What about Privacy?
Each A.I. entity would operate independently of each other yet be capable of communications through a loose network association using the internet as a back bone. This prevents the need for a single server to exist somewhere that downloads ‘updates’ to individual A.I. entities. This should ease the concerns of the privacy advocates to whom this type of technology is alarming and a bit scary. Other safety precautions will have to be put into place to ensure the A.I. is not turned into some sort of oppression tool. These need to take the form of hard coded series of protocols that cannot be over written by someone from the outside. This will need to be true no matter the application.
Multiple devices, one entity
Proper application of A.I. as an entity capable of multiple functions and processes rather than encased in a single use robot would enhance the life of everyone and, if applied properly, will actually lead to a greater and more robust protection of personal information, making ‘identity theft’ obsolete. The Artificial Intelligence entity has an unlimited capacity to learn and, with the same ease as transferring a picture from your phone to your laptop, can be hosted in more than one ‘tool’ or device and freely transferred between them when needed. The best part is the A.I. will already know when it is needed and you can just relax and focus on enjoying life.
One final Question
Are you satisfied that A.I. can be used safely?
At the end of it all...
If we do this right Artificial Intelligence will be the single greatest invention since the telegraph shrunk the world down and allowed us to consolidate knowledge during the Second Industrial Revolution.
But robots alone are not enough.
© 2020 Robin Olsen