So, my previous blog entry was about AI and the trials and tribulations of this computerized wizardry. Why use a term like wizardry? Well, how often was the comedic wizard of comic strip fame or is that comedic infamy... correct? If you, who have often read of that egotistical drunkard, may recall the wizard was most often concocting potions of powder keg disasters.
For the novice who has not read comic strips for a while... a simple meaning of Id is ego. We will not go into a Freudian explanation of interrelated terms of ego.
The purpose of this blog, today, is to prove, and warn, each individual who is concerned about their reputation, that AI will be wrong about many things.
So, I put my novel name, El Dueck, into the search engine and what kind of strange connections did AI make about me and people I'm not related to at all.
Well, if you search for El Dueck, Pinterest, you should come up with the correct person but I may be the only verifier of the myriad photos I have on that website because I am the only person who can recall doing all that work... because I did that work... is that id coming to the fore or is that just the plain truth?
So, the 'online presence' category is correct, today. Will it be for the next 15 years? I don't know.
Now, the 'project mentions' is all wrong because I have no idea who that 'Dueck' is... I don't care how many dots you put in a row.
The 'general quilting' category is also all wrong because I have no idea who those people are either.
So, what can we deduce from this little 'test' of the AI programming 'wizard?' Did the wizard have too much to drink or is the newfound AI overrated and a little dangerous, perhaps?
Allow me to ask you a few questions for the sake of safety. If you had a very common surname, adopted by way too many people,... would you want to be associated with all of them?
Simple answer... of course not! What if there be a criminal in the bunch? Woah! Do not include me then! What if there be a liar in the bunch? Woah! Do not associate me with that one, or two, or thirty... or five hundred.
Now, to prove a point here... if you are not sure yet what the point is... I will simply warn you that the point I am trying to make is a case for or against 'mistaken identity.'
What are the dangers of mistaken identity?
Well, about twelve years ago or maybe fifteen years ago, I was in a hospital and a screen with the names of recent patients scrolled. Factually. What I mean by that is, the people whose names were on the list had all needed a little medical help that week. So, what was so shocking about this list of peoples' names?
Three names were identical... first and last names exactly the same.
Now, this was not random. How can we be almost completely sure that, in this 'hospital' situation, that the names of these individuals were correct but, they were, definitely, not the same three people?
Answer- When needing medical help, I would not want my files all mixed up with another person of a 'similar' name or even an identical name because... each persons' medical history is very individual and must, in this type of situation, be precisely correct, for the safety of patient and health administrators.
So, today, I feel I have proven that AI is not perfect at anything. At times, the connections it makes with you and a person with a similar name may be harmless but at other times, it may be very dangerous.
My best advice to all who use the internet... always use your own intelligence while using the artificial intelligence on the internet because the responsibility of verifying and cross- referencing will always lie with humans,
you and I.
No comments:
Post a Comment