I am a computer science student. I learned about programming languages recently, but I don't know much about artificial intelligence.

I want to know, why don't we program something in a way that we could tell the program

Hey! Do this for me!

And then just sit down and wait that the AI does the job?

Is this currently possible to do?


2 Answers 2


Normally when you write a program, you are acting like a boss that micromanages the job, telling the workers how to accomplish a task, perhaps without even letting them know what the purpose is. What you are hoping to be is a boss that gives the workers a goal and allows them to determine how to accomplish it.

In many ways, that is one of the aims of AI.

We already have small examples, such as a smart washing machine that weighs the clothes, monitors the amount of dirt in the water, and continually decides how much water to add or drain, when to agitate, when to rinse, and when to spin. All you had to tell it was "clean these", and perhaps say what kind of material they are made of.

As a much larger example, automobiles were traditionally operated by turning the steering wheel to cause it to change direction and by pressing the pedals to increase or decrease the speed, but now we are working on car controllers that can respond to "go to Cleveland" by determining the best speed and direction by itself.

But note that those two examples (the first requiring only a little "intellegence", the second a lot) were for very specific requests that could be expressed in a few words. As soon as the request becomes even slightly more complicated, the difficulty of creating a solution becomes much much more difficult.

Ask your "Do this for me!" request of a human being. The first thing they'll do is ask "what does this mean?". And then you'll have to give a lot of details. And then they'll have questions about what you really want. And so on. Providing the requirements for "this" will be neither simple nor easy.

Human intellegence is still vastly superior to current AI. In particular, it is capable of not only recognizing that additional information is required, but of having the intuitive ability to know what that missing information must be like.

  • 1
    $\begingroup$ Not writing my own answer because this hits many of the points I wanted to make, but I feel you're missing an important one. AI at the moment can not infer - it has no intuition. If you give it a complicated set of instructions, unless it has immense amounts of experience with similar instructions, the AI will fail miserably. An AI that could adequately decipher complicated instructions would (in my opinion) be self aware, so until we reach that point, I don't think a general AI for performing actions from complicated instructions can be a thing. $\endgroup$
    – Recessive
    Jun 24, 2020 at 6:59
  • $\begingroup$ @Recessive, I've expanded the answer. If I missed the point you wanted, feel free to edit the answer. $\endgroup$ Jun 24, 2020 at 18:55

This is still a long way off and would require what is known as Artificial General Intelligence, which is probably the best descriptive term for AI with cognitive capacity akin to humans.

(Another term often used is "superintelligence", but it's less precise in regard to general intelligence, in that it could be understood to connote superior intelligence that is not general. The same problem exists for Searle's "Strong AI", an AI that matches or exceeds humans capacity, since AI can now be "strong but narrow", exceeding human capability in a single task or a set of tasks, such as a game, or managing fully definable systems such as air-conditioning.)

Part of the problem is we still don't really understand how human cognition works, so the famous projection by an incredibly important figure in AI, Herbert A. Simon, that:

"Machines will be capable, within twenty years, of doing any work a man can do."

is still, like fusion power, potentially always 20 years in the future. (Worth noting that if a general superintelligence is developed, it could not only do anything humans could do, but things we can't even conceive of, because its intelligence would greatly exceed our own, and would likely be capable of continual self-optimization.)


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .