The editing of the robot column did not differ from the editing of a human article, but took less time, the newspaper reported.
Neural network GPT-3, developed by the AI company OpenAI (one of its founders – Ilon Mask), wrote a column for the British newspaper The Guardian. Its headline is as follows: “This article was fully written by the robot. Are you already scared, human?”. The material is published on The Guardian website.
The author was tasked to create an essay of about 500 words about why people should not be afraid of AI.
“The journalists wrote only a small introduction to make it easier for the robot to navigate,” explained the publisher.
The article says that the AI has no purpose to harm humanity. Also, the author claims that he is a friendly robot and a servant of people, although they do not trust him. The “Columnist” urges readers to trust computers more because the main task of technology is to make people’s lives easier and better. The algorithm added that, on the contrary, it would even sacrifice itself for humanity. He also added that he does not want omnipotence – it is simply not interesting, as well as all kinds of violence.
It is reported that the algorithms gave a total of eight versions of the article, which then editors combined into one material.
“It wasn’t heavier to correct it than a human text – it was even easier,” said The Guardian.
The algorithm ended his column with a quote from Mahatma Gandhi: “A small body, driven by the spirit and inspired by an unquenchable faith in its mission, can change the course of history.
“I can do it,” she added at the end of GPT-3.
GPT-3 is a new continuation of the GPT-2 neural network. Like its predecessor, it is able to continue text in any format and style. Only now the neural network is more powerful and can create more linked texts.
Earlier Vancouver programmer Murat Eifer developed an artificial intelligence capable of thinking about the meaning of life and answering philosophical questions.