Test its possibilities.
Originally discussed, the AI was to be used for answering questions, describing stories or translating texts. However, Open AI researchers quickly came to the conclusion that their algorithm could be used for quickly filling the network with fake news.
OpenAI decided to give the world a full-fledged edition of its artificial intelligence, as its simplified version did not lead to abuse on a larger scale, and basically just because of that. The organization hopes that its algorithm will help scientists from around the world to develop better models for detecting text generated by artificial intelligence and to eliminate language bias from language models.
It should be noted that the model GPT-2, because that is the name of artificial intelligence, still has the ability to generate reliable fake news – it has not changed. This means that anyone can use it at any time to spread propaganda based on fake news. Although OpenAI has not yet seen abuse on a larger scale, in my opinion it is only a matter of time.
Interestingly, each of you can test this artificial intelligence from OpenAi. This possibility is guaranteed by the website named talktotransformer.com. Just go to it, enter any text in the visible box (or choose one of the suggested ones), and then click on the phrase "complete text". It is best if you enter the text in English. Reaching for a phrase written in Polish gives very funny results. Apparently the GPT-2 model was not created with the intention of generating texts mainly in Polish.