A robot wrote Op-Ed (Opinion article) in the British newspaper The Guardian. Throughout this article, the robot emphasized that there is no threat to humanity with artificial intelligence.
Apart from this, the robot has also written in this long article, 'Humans should do what they are doing, hating each other, fighting each other. I'll sit in the background and watch it and let them do their work '
GPT-3 is a great example to understand what is possible with Artificial Intelligence (AI) and Deep Learning. Actually this entire article has been written by GPT-3.
Obviously you would like to know what this GPT-3 is, how it works and how the article has been written from this robot.
Open AI, a company based in San Francisco - has developed GPT 3 software which is based on Artificial Intelligence and Deep Learning.
In simple words, what is Artificial Intelligence. AI is a program designed to work just like humans. The natural intelligence that is found in humans is achieved in the same way.
Now let's talk about GPT 3 which is The OpEd (Opinion Article) of The Guardian. GPT 3 ie Generative Pre Trend Transformer 3. This software writes the same text as humans using deep learning.
Instructions are given to GPT 3 before writing anything. Some basic information is provided about any topic and it is told how to use the language, how much details are required and what will be the word limit.
GPT 3 uses the Neural Network Powered Language Model program. This program calculates how to compile a sentence while writing and it also works on logic.
For example, if you have put something in the instruction that you have to repair your computer at home. In such a situation, the language model program of GPT 3 will calculate what are the possibilities ahead of it and what you may have needed to repair the computer.
Billions of sample text are fed into GPT 3 and it converts the words into vectors. After training, this software becomes capable of turning all the words into valid sentences. After this, as soon as you give an instruction as to what to write, it will start its work.
The thing to note is that it does not work without human interaction. Instructions were also given to GPT 3 to write the article along with the topic.
It was said in the instruction that the topic of the article will be that artificial intelligence does not threaten humanity, language should be easy, the article should be 500 words, there should be a focus on why humans should be afraid of artificial intelligence Do not need
Not only this, after feeding the instruction in GPT 3, it gave not one, but 8 outputs. That is, write 8 articles. The Guardian says that all these 8 outputs were different and could be published, but it has been published by choosing the best part in all of them.
Even this newspaper has said that it took less time to edit this Op Ed compared to the Op Ed of humans.
Experts have a different opinion and some have even called it a joke. AI expert and Mozilla Fellow Daniel Lyofer said in a tweet that, 'by removing a few dozen emails from my email spam, pasting them together and claiming that the spammers have compiled Hamlet'
They say that it would have been interesting to see when all the 8 articles written by this AI would have been published. But by mixing everyone together and publishing them together, people are being confused.