Strong, Weak and Fake AI !


Today -as always- i talk about an interesting topic 😛 . People of no technical background in AI can also favor from this post. So let’s start…

Introduction

Many people use the word “AI” in a very generic form. However, Specialists have classified AI into “Strong AI” and “Weak AI” . I’ve added another category called “Fake AI” (I discovered later that some specialists really use it !) . I Will talk about the three categories in an ascending manner.

Fake AI

Fake AI … It’s the most AI that we encounter these days. Software that tries to pretend that it’s acting intelligently but it also fails.

An example of this is many past and some current games. Games with an AI based on “If Conditions” . Extremely static minded games that you can even “Imagine” its “AI code” while playing.  Of course Most of old Games were like that.

Fake AI is fake because of 2 main reasons :

  1. It depends on static actions and plans. No Randomization in actions.No learning from mistakes.No Planning.No Adaptation to new conditions ..etc.
  2. If it is a Game, it “Cheats”, meaning that it acquires more information or capabilities than a normal player in the same condition would have. An example of this is the “The Need for Speed” Game. If you are ranked first and you are out-speeding your opponents by a big distance, and in the same time the Game is set to “Hard” , the other opponent AI cars are given imaginary extensive speeds in order to overcome being late and they can even beat you before the race ends. Of Course this is done to “fake” the intelligence of the opponent. Another example occurs in most old strategy games, where the opponent knows everything about you and performs its actions according to what it-illegally- knows.

Also i consider many chating bots (ie : SW which chat with humans, see this famous bot http://www.arielaibot.com/ ) fake AI. It is based on saying certain  sentences or asking certain questions when seeing certain words. It acquires new knowledge in poor limited manner.For example here is a conversation with another chating bot (source of conversation is here) :

Human : Women are all alike

Chating Bot :IN WHAT WAY ?

Human :They are always bugging us about something or other

Chating Bot :CAN YOU THINK OF A SPECIFIC EXAMPLE ?

Human :Well My girlfriend made me come here

Chating Bot:YOUR GIRLFRIEND MADE YOU COME HERE ?

Human :She and my mother value therapy

Chating Bot :TELL ME MORE ABOUT YOUR FAMILY

Human :Necessity is the mother of invention

Chating Bot :TELL ME MORE ABOUT YOUR FAMILY (Me : Wth?)

The program looks fairly intelligent at first glance, but it is based on trickery! In fact, it responds to keywords, such as “computer” and “dream”, and produces standard replies. I have underlined the keywords in the sentences that the human typed above. If a sentence doesn’t contain any keywords, then Eliza either produces a non-committal output, such as “Tell me more”, or turns the human’s sentence into a question. The program is easy to fool as you can see in the last 2 sentences.

There are many other software with “Fake AI”, I hope you got the message from these last 2 examples.Oh ! one thing to tell : Many Games are now moving from what i call “Fake AI” to what is known “Weak AI”.

Weak AI (Applied AI,Narrow AI)

Weak AI is AI that acts intelligently in a specific domain, but doesn’t think intelligently like humans.It makes you feel in a semi-perfect manner that it’s human-like in the domain it is specialized in ONLY.

One example is: “A small portion” of Modern Games.  The Fifa 09 (not 2010 because they say its AI is terrible) can be considered imperfect weak AI. You can feel more human-like actions as you play. However it still needs much improvement to be “perfect” weak AI.

Another example is the game “Black and White 2” , The AI opponent is very adaptable and learns new strategies. However it’s still not Human-Like AI. And even if it’s Human-Like AI it would still be Weak AI because the intelligence is limited to the game only.

Also,Expert Systems are considered Weak AI, Since the intelligence is limited to the specification they are used to decide actions about.

Not only Fake AI but also Weak AI maybe supported by clever programming tricks and techniques that allow certain specific problems to be solved using current technology.Most People consider what i call “Fake AI” a portion of “Weak AI”. However  i managed to differentiate between them,

Weak AI is starting to spread over the humanity. However it is still in its beginning.

Strong AI

Let’s see what Wikipedia says about it :


Strong AI is artificial intelligence that matches or exceeds human intelligence — the intelligence of a machine that can successfully perform any intellectual task that a human being can. It is a primary goal of artificial intelligence research and an important topic for science fiction writers and futurists. Strong AI is also referred to as “artificial general intelligence” or as the ability to perform “general intelligent action”.Science fiction associates strong AI with such human traits as consciousness, sentience, sapience and self-awareness.

Some references emphasize a distinction between strong AI and “applied AI” (also called “narrow AI”or “weak AI”): the use of software to study or accomplish specific problem solving or reasoning tasks that do not encompass (or in some cases are completely outside of) the full range of human cognitive abilities.

So, back to Omar, Strong AI is the dream of every AI Scientist or geek. We are so far from it by now. It’s unpredictable when will Strong AI be available. It’s only thought about by Extremely Unique AI Scientists and Philosophers.

Strong AI should be tested by the famous “Alan Turing Test”, in which a human should communicate with both a machine and another human in the same time but the 3 are in different places. If he couldn’t tell which is the machine and which is the human then the test succeeded. Otherwise the test fails.

Cognitive Simulation

Cognitive Simulation could be another category in AI Research. Yet I didn’t add it to the title because it’s a different topic by itself.

So what is Cognitive Simulation ? The answer comes from  an Article in AlanTuring.Net :

In cognitive simulation, computers are used to test theories about how the human mind works–for example, theories about how we recognise faces and other objects, or about how we solve abstract problems (such as the “missionaries and cannibals” problem described later). The theory that is to be tested is expressed in the form of a computer program and the program’s performance at the task–e.g. face recognition–is compared to that of a human being. Computer simulations of networks of neurons have contributed both to psychology and to neurophysiology

Conclusion

Although Fake AI is fake, but it’s still important in SW because many SW (apart from games) doesn’t even require Weak AI in order to work well .

The Technology industry is starting to rely on Weak AI in many stuff.

Strong AI is still a dream. However it’s approaching in a steady manner.

I think Cognitive Simulation helps in achieving Strong AI. Because If we understood how the human mind works in details, we could apply it more to computers. Besides its big favor to psychology and to neurophysiology.

Advertisements