Hard AI vs Soft AI

AI (Artificial Intelligence) is “the study and design of intelligent agents, where an intelligent agent is a system that perceives its environment and takes action, which maximises its chance of success” (Wikipedia).

Throughout the web, there are AI systems that are able to mimic human responses, leading the recipient to believe that there are talking to someone else (i.e. through social-networking sites) when they are actually communicating with computing intelligence. This is in fact known as “hard AI”, this creates machines that are made to think like us in order to understand how the human brain works, in a result whenever we communicate with AI we tend to think we care talking with another human being. Many sites however use security checks (CAPTCHA) when making a social-networking site or new e-mail account (such as Facebook and Hotmail) so that these AI cannot make fake accounts, as they cannot see the text but just hidden code.

‘Eliza’ is a computer program that tries to communicate in human form. Due to it being very basic programming, Eliza is extremely limited in terms of responses, only picks up on a few key words and doesn’t give a straight answer but tends to answer a question with a question.

Another similar type of AI programming is ‘A.L.I.C.E’, which is far more developed in terms of thought structure and explaining questions with impressive detail. It is far more human like as it uses ‘smiley faces’ which would leave us to believe that it was another human as smiley faces represents emotions.

Eliza first appeared in the 1960’s and A.L.I.C.E was connected to the net in 1995. The advancements it took from the basic program of Eliza has been impressive as A.L.I.C.E was created 35 years later. This shows that technology and AI advancements are on the rise. The question that arises would be if they are able to create an AI that is superior to its creator meaning it will become human like in every way such as showing emotion or having conversations with humans beyond the intellectual level but on something that will differ a robot from a human or if there is a limit on how human like they can make a robot.

I think that robots will always be robots, as they can only read texts that are put in, they are unable to get around the (CAPCHA) security and therefore unable to read text that humans can read, therefore limiting them. Although they can gives responds that are human like, they cannot give genuine opinions that they have thought of as they only come from the person that created them. Then again in another 35 years this may all be possible.


A.L.I.C.E
Eliza
AI deffinition