An Unfinished Artificial Intelligence Idea

Learning program

When it is invented, will make more money than anything. Because then robots will be able to perform complex tasks for us(cooking, cleaning, fixing, building). And will be able to be friends to us.

Communication: Robots would communicate to each other through tones in the same way that “R2D2” does. This is a much easier to decode language that would be captured by hearing devices. And may be translated to us through a translator earphone. And we can also talk to them or ask them a question. This may require a new, and less complex language than english for a robot to understand questions and sentences.

The way to make a learning robot is cause and effect and to learn from mistakes. Positive reinforcement and negative reinforcement.

Questioning, as a means to learn the universe around them: To fill in blanks left in the robots universe, it should ask us, Who what when where why. Because we don't know what it knows or doesn't know unless it asks.

When they know that gravity effects objects then they know to aim above an object they are throwing a ball at, in a parabolic arch.

Vocabulary: They must have a vocabulary to be able to describe the universe around them. Each word is given a tonal code in order for us to tell it that work in tones and for it to translate those tones into the word.

The robots must always ask why about everything, and store the answer, to know the answer to that question later on in life.

Vision: must be just as good as ours and they must be able to know what they are seeing and also predict that objects actions. But vision may not be necessary for a mostly talking robot. Also objects can have Radio Frequency Identification (RFID) Chips on them so that the robot can know what the item is and that items properties so that it knows how to handle that item. And later on it can learn what the item looks like and not even need the RFID chips at all to recognize something.
Also you need a "spacial actualization program" in the brain knowing where an object is and what shape it is and how big it is in relation to you. That's where 3-D vision comes in handy. If your robot sees a stuffed animal they need to be able to recognize what it is without turning it around which means being able to fill in gaps of information with assumptions. Magic isn't possible also so the robot needs to differentiate between plausible explanations of what the object is and reject implausible explanations. For instance filling in the blanks when asked "What is in the front yard right now", it's usually not answered with "a pink elephant".
And they need a way to differentiate between different shapes and be able to make up novel descriptions of newly shaped objects attributes that they come across. And be able to differentiate the relevance of different attributes of an object. Example: To recognize a teddy bear we see it's facial features and shape and texture and we know that is is a teddy bear. So describing physical attributes is key to being able to recognize objects.

Neural memory banks: And the robot would tell you when its memory banks were full so that you could buy them another flash memory drive.

And we could copy their “mind” and download it onto a second clone robot.

Use it or loose it. When the robot hasn’t done something for a while it will forget (or "erase" if you will) that information making them have to problem solve again in order to find a new way of doing that same thing. Sort of like pushing the reset button on a video game. In order to come up with new novel approaches to solving problems.

Problem solving: Is done by having a goal and finding a way to reach that goal, which will not need second chances to accomplish. And if the attempt at reaching a goal fails you must consider risk and reward. If the risk is someone dieing then don't even attempt trying to reach your goal unless absolutely sure of success.
Problem solving, and real world based critical thinking, is key for original thought processes or sentience.

And imitation is key to passing, already perfected ways of doing things, onto new generations for a head start in life.

Eventually after a mass produced copy of an already advanced working learning robots mind is made it will be called a "personality chip" as a more specific term than just the generic term "artificial intelligence chip".