Rise of the Machine: Artificial intelligence in the author's seat, not the future but the present
 |
| Rise of the Machine: Artificial intelligence in the author's seat, not the future but the present |
One question for you as a reader is, if I were to tell you that this article is not the work of a human author, but of an artificial intelligence, or more simply a computer program, how would you feel when you read it? Like reading a man-made text, or more exciting? Or will the writing lose credibility due to the creation of an emotionless device?
So far, there has been no doubt that any writing on any subject is man-made. However, at present, there is a possibility that the matter may not be completely true. Let's start with the closest example.
One of the most popular features in today's smartphone keyboard applications is 'Next Word Suggestion.' When writing a sentence, artificial intelligence guesses what the next word might be after writing a word. Displays more than one possible word on the screen, including the highest possible word by guess. A lot of times, that assumption matches.
 |
| Next Word Suggestion |
When it comes to AI or artificial intelligence, the names Google Assistant, Siri, or Amazon Alex come to mind. But this time, let's introduce another new artificial intelligence. Producer Adam King has named Talk to Transformer. Before reading its features, let's read the following part carefully.
It was cold inside the house. Darkness seemed to be swallowing me, an unknown fear was playing in my thoughts. I felt like I was not alone in the house. Then it happened, suddenly a cold wave flowed down my spine. And then I felt like I was "walking on air", even though I'm actually frozen. I was flying."
My brain was unable to process what was happening, which is why I knew that I must have had a stroke or some sort of temporary death, but I didn't know how long it would be. The thing was so bizarre and so different from the whole state of mind that I had just started to feel a strong physical sensation in my hands. I was trying to look at the outside world, but nothing was….
The above paragraph is not part of any supernatural novel. After inputting the initially bolded part to 'Talk to Transformer,' the rest of the part is written by the AI as its own. That is, people have tried to imitate writing without the use of any template, even if it is not collected from anywhere. Just reading a part of the story, did you understand that the writing is not completely man-made?
Language modeling is the technology used behind 'next word suggestion' or 'text generating' artificial intelligence. The more advanced model used in the case of 'Talk to Transformer' is the language model GPT-2 made by OpenAI Inc. However, in these cases, it is much easier for the models to write 'fiction' than to create 'fact.'
Whether or not Artificial Intelligence will sit next to a writer and write the next 'Harry Potter' or 'Lord of the Rings' series of fiction, it's easy to say what this language model is all about before going into the debate, and how-or AI could write the rest on its own. Let's take a little idea of the language. We will then try to determine whether artificial intelligence is going to be our comrade-in-arms in the pen war, or an invincible opponent, or something other than the pen war.
The language modeling method has taken 'natural language processing' to another height. One of the branches of artificial intelligence is natural language processing, which is how computers communicate in human language. This branch deals with technologies such as talking to Google Assistant, translating from one language to another, and converting text to speech.
The language model works based on probability and statistics. It is programmed in such a way that the more one word is used with another, the more artificial intelligence can guess.
How powerful the AI will depend on how much information it is being 'trained' with. The more information provided, the more it will learn, and the less it will make mistakes. This requires a lot of accurate information. For example, man-made translations from one language to another or any kind of email or message spam for spam filters are being identified. This information gradually reduces the number of errors by analyzing artificial intelligence.
But writing a whole story or paragraph does not require any kind of information. In other words, if you want, the language model can learn the language by analyzing different stories or articles from thousands of books or millions of websites. By analyzing millions of articles on Wikipedia, one can understand which word is more related to which word. The language model creates a mathematical map of word possibilities and statistics. In the language of artificial intelligence, it is called 'Self Supervised Learning.' With the addition of some complex algorithms, this model can reach a different level with words. One of the easiest tasks is to create small fictional paragraphs on any topic.
At the end of 2016, an organization called Botanic Studios introduced its artificial intelligence to J.K. Rolling's seat. Analyzing all the novels in the Harry Potter series, he wrote another story in AI, entitled 'Harry Potter and the Portrait of What Looked Like a Large Pile of Ash.' The three-page book with this ridiculous title provided food for laughter, but according to the reader's response, the writing of artificial intelligence was not a bad one. However, this skill of artificial intelligence is not limited to making a part of the story, has created a book of poetry, has also written news and articles.
In 2016, data scientist Ross Goodwin ran a project with the help of Google. He drove from Brooklyn to New Orleans in a black Cadillac car with a GPS system, a special camera, and a microphone. During the journey, the three devices receive information about the surrounding environment and send it to the neural model inside the laptop. Poetry is written in keeping with the surrounding environment and time. At the end of a four-day-long journey, the first book of poetry written in history, 'One the Road,' was written by artificial intelligence. Although not man-made, it lacked artistry. Paper copies of the book are also available for purchase.
Based on a language model called Grover, Adam Getz has created a website called ' News You Can't Use,' which contains only fake news, all of which is written by artificial intelligence. This model has been trained with news articles from more than 5,000 publications. Interestingly, the model did not maintain the balance of information from the beginning to the end, only imitating the way people write the news.
The movie script is being written using the language model. Many more such examples can be given to prove the achievement of artificial intelligence. But does that prove that artificial intelligence will enter the competition of this creative talent with people in the future? Will the true writer lose his place, just as humans lose their jobs to robots? Or is our fear of innovation completely unfounded?
There is an important thing at the root of so many stories that have been narrated from the beginning, which is information or data. Everything that the machine does, whatever it decides, is all based on data analysis. Artificial intelligence does not create any new information by thinking anew here. A story or a novel is not just a rearrangement of sentences consisting of a few words. Each novel is simultaneously an expression of experience, expression, and novelty. An inexperienced, inexperienced computer that is still not quite proficient in the subject. Although experienced in working with information, these computer programs do not know how to apply creative thinking.
Another big part of the limitations of the current language model is authenticity verification. What these technologies are writing cannot determine for itself how logical it is. Artificial intelligence can never replace people by writing novels in a combination of unrealistic subjects. At the same time, maintaining important writing consistency. The range of subjects written so far is very small, yet it has been observed that there is considerable inconsistency in those writings. No artificial intelligence has yet emerged to write a large-scale novel consistent from beginning to end. So in the current context, this technology is not only a competitor but also a great tool that can teach people to think from a different point of view. However, many researchers believe that it is possible to overcome these complications shortly. Will the author then displace people, the way jobs for robots are disappearing at the present time?
One thing to keep in mind here is that robots or artificial intelligence are not putting people in a job crisis, but the authorities are cutting workers in the hope of making a profit. Future publishing companies must calculate the profit margin very well, as the computer does not have to pay dividends for any AI-generated book.
However, in the current context, there is a bigger concern with this technology than writing a novel, that is bias. In the hope of neutral results, we have turned to technology. But does artificial intelligence really play a neutral role?
According to a report published by Pro-Public in 2016, an AI program 'Compass' used in a US court has revealed the possibility of re-incarceration of prisoners. Surprisingly, however, the program mistakenly shows that black offenders have a higher crime rate each time than whites. When asked how the program works, the manufacturer declined to comment. Similarly, another program, Predpole estimates, which area has the highest crime trend; According to him, police security has been strengthened there from the east. This time too, the result is that there is a high risk of crime being committed in the homes of blacks.
In both cases, the programs are trained with information on previous police reports, crime cases, etc. In other words, programs are taught exactly what people have come up with so far, not neutrality. It has been said before that what kind of information is used to train these programs is very important. The results will be exactly the kind of information we provide. We can never expect neutrality with repetitive information from our own history. The same can be said about writing?
With these language models, it is very easy to create fake news, writing on controversial topics or news that can hurt racism or religious sentiments. The Facebook-Cambridge Analytica scandal has pointed the finger at us, how to calculate the cost of applying technology in the wrong place. It is not difficult to imagine that the use of these technologies in the wrong hands is one of the weapons of political propaganda.
Nowadays, where fake news is circulating across the internet, artificially made fake news is like adding fuel to the fire. The technology can also involve fake product reviews or arguments with people online.
However, the important question now is - who is responsible for controlling these newly created content? Who owns the rights or responsibilities? Most of the laws that are in place do not clearly prescribe anything for these technologies. Many times the development of technology from the law is fast. Over time, the workings of technologies have become more complex, making it difficult for the public to understand. Manufacturers are also not showing the transparency of their technology.
Technology like Google Duplex has made us think about the users on the other side of the screen for the sake of technology. Artificial intelligence has mimicked human appearance, voice; And now the language is doing. The existence of this artificial intelligence is making us think anew where human beings are not a safe place online.
This article discusses not only the role of artificial intelligence as a writer but also the fears that arise in our minds, but in all cases, artificial intelligence should not be viewed in a negative light or as a cause of fear. Rather, a group of people is working tirelessly with this technology to make our lives easier, with a lot of respect and love for them. What do you think? Be sure to let me know. How do you see this technology in the future?
Language modeling is the technology used behind 'next word suggestion' or 'text generating' artificial intelligence. The more advanced model used in the case of 'Talk to Transformer' is the language model GPT-2 made by OpenAI Inc. However, in these cases, it is much easier for the models to write 'fiction' than to create 'fact.'
Whether or not Artificial Intelligence will sit next to a writer and write the next 'Harry Potter' or 'Lord of the Rings' series of fiction, it's easy to say what this language model is all about before going into the debate, and how-or AI could write the rest on its own. Let's take a little idea of the language. We will then try to determine whether artificial intelligence is going to be our comrade-in-arms in the pen war, or an invincible opponent, or something other than the pen war.
The language modeling method has taken 'natural language processing' to another height. One of the branches of artificial intelligence is natural language processing, which is how computers communicate in human language. This branch deals with technologies such as talking to Google Assistant, translating from one language to another, and converting text to speech.
The language model works based on probability and statistics. It is programmed in such a way that the more one word is used with another, the more artificial intelligence can guess.
 |
| Language model sound estimation; Image Source: MLexplained |
How powerful the AI will depend on how much information it is being 'trained' with. The more information provided, the more it will learn, and the less it will make mistakes. This requires a lot of accurate information. For example, man-made translations from one language to another or any kind of email or message spam for spam filters are being identified. This information gradually reduces the number of errors by analyzing artificial intelligence.
But writing a whole story or paragraph does not require any kind of information. In other words, if you want, the language model can learn the language by analyzing different stories or articles from thousands of books or millions of websites. By analyzing millions of articles on Wikipedia, one can understand which word is more related to which word. The language model creates a mathematical map of word possibilities and statistics. In the language of artificial intelligence, it is called 'Self Supervised Learning.' With the addition of some complex algorithms, this model can reach a different level with words. One of the easiest tasks is to create small fictional paragraphs on any topic.
At the end of 2016, an organization called Botanic Studios introduced its artificial intelligence to J.K. Rolling's seat. Analyzing all the novels in the Harry Potter series, he wrote another story in AI, entitled 'Harry Potter and the Portrait of What Looked Like a Large Pile of Ash.' The three-page book with this ridiculous title provided food for laughter, but according to the reader's response, the writing of artificial intelligence was not a bad one. However, this skill of artificial intelligence is not limited to making a part of the story, has created a book of poetry, has also written news and articles.
 |
| How intelligent is artificial intelligence? |
In 2016, data scientist Ross Goodwin ran a project with the help of Google. He drove from Brooklyn to New Orleans in a black Cadillac car with a GPS system, a special camera, and a microphone. During the journey, the three devices receive information about the surrounding environment and send it to the neural model inside the laptop. Poetry is written in keeping with the surrounding environment and time. At the end of a four-day-long journey, the first book of poetry written in history, 'One the Road,' was written by artificial intelligence. Although not man-made, it lacked artistry. Paper copies of the book are also available for purchase.
Based on a language model called Grover, Adam Getz has created a website called ' News You Can't Use,' which contains only fake news, all of which is written by artificial intelligence. This model has been trained with news articles from more than 5,000 publications. Interestingly, the model did not maintain the balance of information from the beginning to the end, only imitating the way people write the news.
The movie script is being written using the language model. Many more such examples can be given to prove the achievement of artificial intelligence. But does that prove that artificial intelligence will enter the competition of this creative talent with people in the future? Will the true writer lose his place, just as humans lose their jobs to robots? Or is our fear of innovation completely unfounded?
There is an important thing at the root of so many stories that have been narrated from the beginning, which is information or data. Everything that the machine does, whatever it decides, is all based on data analysis. Artificial intelligence does not create any new information by thinking anew here. A story or a novel is not just a rearrangement of sentences consisting of a few words. Each novel is simultaneously an expression of experience, expression, and novelty. An inexperienced, inexperienced computer that is still not quite proficient in the subject. Although experienced in working with information, these computer programs do not know how to apply creative thinking.
Another big part of the limitations of the current language model is authenticity verification. What these technologies are writing cannot determine for itself how logical it is. Artificial intelligence can never replace people by writing novels in a combination of unrealistic subjects. At the same time, maintaining important writing consistency. The range of subjects written so far is very small, yet it has been observed that there is considerable inconsistency in those writings. No artificial intelligence has yet emerged to write a large-scale novel consistent from beginning to end. So in the current context, this technology is not only a competitor but also a great tool that can teach people to think from a different point of view. However, many researchers believe that it is possible to overcome these complications shortly. Will the author then displace people, the way jobs for robots are disappearing at the present time?
One thing to keep in mind here is that robots or artificial intelligence are not putting people in a job crisis, but the authorities are cutting workers in the hope of making a profit. Future publishing companies must calculate the profit margin very well, as the computer does not have to pay dividends for any AI-generated book.
However, in the current context, there is a bigger concern with this technology than writing a novel, that is bias. In the hope of neutral results, we have turned to technology. But does artificial intelligence really play a neutral role?
According to a report published by Pro-Public in 2016, an AI program 'Compass' used in a US court has revealed the possibility of re-incarceration of prisoners. Surprisingly, however, the program mistakenly shows that black offenders have a higher crime rate each time than whites. When asked how the program works, the manufacturer declined to comment. Similarly, another program, Predpole estimates, which area has the highest crime trend; According to him, police security has been strengthened there from the east. This time too, the result is that there is a high risk of crime being committed in the homes of blacks.
In both cases, the programs are trained with information on previous police reports, crime cases, etc. In other words, programs are taught exactly what people have come up with so far, not neutrality. It has been said before that what kind of information is used to train these programs is very important. The results will be exactly the kind of information we provide. We can never expect neutrality with repetitive information from our own history. The same can be said about writing?
With these language models, it is very easy to create fake news, writing on controversial topics or news that can hurt racism or religious sentiments. The Facebook-Cambridge Analytica scandal has pointed the finger at us, how to calculate the cost of applying technology in the wrong place. It is not difficult to imagine that the use of these technologies in the wrong hands is one of the weapons of political propaganda.
Nowadays, where fake news is circulating across the internet, artificially made fake news is like adding fuel to the fire. The technology can also involve fake product reviews or arguments with people online.
However, the important question now is - who is responsible for controlling these newly created content? Who owns the rights or responsibilities? Most of the laws that are in place do not clearly prescribe anything for these technologies. Many times the development of technology from the law is fast. Over time, the workings of technologies have become more complex, making it difficult for the public to understand. Manufacturers are also not showing the transparency of their technology.
 |
| Rise of the Machine and Artificial intelligence |
Technology like Google Duplex has made us think about the users on the other side of the screen for the sake of technology. Artificial intelligence has mimicked human appearance, voice; And now the language is doing. The existence of this artificial intelligence is making us think anew where human beings are not a safe place online.
This article discusses not only the role of artificial intelligence as a writer but also the fears that arise in our minds, but in all cases, artificial intelligence should not be viewed in a negative light or as a cause of fear. Rather, a group of people is working tirelessly with this technology to make our lives easier, with a lot of respect and love for them. What do you think? Be sure to let me know. How do you see this technology in the future?
Hasibul
I love blogging. If everyone is with me, I can take this blog further. For this I ask all of you for help and prayers.
