Recently, a well known non-profit OpenAI released an AI system (GPT-2) capable of producing texts of such high quality that most readers will be convinced it was written by a human.
By feeding GPT-2 a couple of sentences it can continue completing the initial text on its own, producing stories of impressive quality.
To illustrate let us look at an example from OpenAI, these initial sentences were written by a human:
‘In a shocking finding, scientist discovered a herd of unicorns living in a remote, previously unexplored valley, in the Andes Mountains. Even more surprising to the researchers was the fact that the unicorns spoke perfect English.’
The AI then proceeded and after 10 attempts wrote the following continuation of the story:
‘The scientist named the population, after their distinctive horn, Ovid’s Unicorn. These four-horned, silver-white unicorns were previously unknown to science.
‘Now, after almost two centuries, the mystery of what sparked this odd phenomenon is finally solved. ‘Dr. Jorge Pérez, an evolutionary biologist from the University of La Paz, and several companions were exploring the Andes Mountains when they found a small valley, with no other animals or humans.
Pérez noticed that the valley had what appeared to be a natural fountain, surrounded by two peaks of rock and silver snow.
For full story that AI wrote (it is long) please check the end of the blog post. It is impressive.
On the technical level, the transformer architecture used in the model seems to be one of the big leaps in AI and is bettering the widely known LSTM model.
OpenAI decided not to release the language model’s code and training set (opting for a smaller model) due to fears that it will be abused for generation of fake stories and product reviews among others. But the paper on which GPT2 is based is available:
OpenAI also released the smaller model, available here: https://github.com/openai/gpt-2 .
Even though OpenAI took a cautious approach and rightfully so, as with all technology it is only a question of when somebody repeats the results and deploys it. We ourselves are already working on similar transformer models and you can use some of them on our platform.
What are the implications for various fields?
Journalism is certainly one that could be even more under threat than it already is. GPT-2 like systems would probably perform well in writing stories when given the initial facts. And as the following story shows even GPT-2 is already impressive at this task:
But the field which is nearer to us is content generation and online marketing. We all know that great content is one of the essential paths of attracting visitors and backlinks.
So the interesting question is: as better versions of GPT-2 from OpenAI or others become available, could such an AI write absolutely stunning pieces of online marketing content, bettering even the best humans?
And will online marketing companies have to take a serious look at AI models to keep up the pace in the next decade just because the competition will be looking and using this?
We firmly believe that systems like GPT-2 will radically transform the content generation industry:
- they will dramatically reduce the price to produce outstanding content
- they will provide us with content of impressive quality
- initial models will be specialised for fields, e.g. writing for online travel industry, products reviews, etc. This is also our development path.
- rise of new fields and niches (emergence of AI quality editors, equivalent to proof editing of neural machine translation)
- combination with neural machine translation (NMT) will enable generation of content for almost all languages (we already employ NMT in our platform)
We aim to be at the forefront of AI content. Become our client and join us.
All our clients who purchase the annual Gold Partner plan will get our current and future AI content generation solutions for free at 500 API calls per month (and with 50% discount for higher numbers of API calls).
‘Pérez and the others then ventured further into the valley. “By the time we reached the top of one peak, the water looked blue, with some crystals on top,” said Pérez.
‘Pérez and his friends were astonished to see the unicorn herd. These creatures could be seen from the air without having to move too much to see them – they were so close they could touch their horns. While examining these bizarre creatures the scientists discovered that the creatures also spoke some fairly regular English. Pérez stated,
“We can see, for example, that they have a common ‘language,’ something like a dialect or dialectic.” ‘Dr. Pérez believes that the unicorns may have originated in Argentina, where the animals were believed to be descendants of a lost race of people who lived there before the arrival of humans in those parts of South America.
‘While their origins are still unclear, some believe that perhaps the creatures were created when a human and a unicorn met each other in a time before human civilization.
According to Pérez, “In South America, such incidents seem to be quite common.” ‘However, Pérez also pointed out that it is likely that the only way of knowing for sure if unicorns are indeed the descendants of a lost alien race is through DNA. “But they seem to be able to communicate in English quite well, which I believe is a sign of evolution or at least a change in social organization,” said the scientist.’