Center for Strategic Assessment and forecasts

Autonomous non-profit organization

Home / Science and Society / New in Science / Articles
"It is more dangerous than any weapon". The creators of the new AI coded their development
Material posted: Publication date: 30-08-2019
American company OpenAI development in the field of artificial intelligence, introduced a new version of self-learning neural networks GPT-2. The algorithm of AI was so impressive that the creators had to disable part of its functions, and journalists have compared its potential threat with real weapons.

the company, the OpenAI created a new version of the generator of texts with the potential to create an industrial scale to create fake news and dangerous spam with effective clickbaiting headers. Learning system GPT 2 developers used the eight million web pages. Allegedly, it is able to adapt to the style and content of a given text fragment, for example, to finish a poem Shakespeare or write a note using one initial offer.

The company representatives said that they were forced to disable some functions of the system before the announcement because of the possibility of its use in malicious ways. Journalists were able to test the "stripped-down" system, after entering one of the proposals they received from artificial intelligence is quite coherent meaningful text, even though it was a glimpse of "native writing".

"It works well enough to use it as a General artificial intelligence to generate text. I'm terrified of GPT-2 because it is a technology that can be used to manipulate the population. In my opinion, that makes him more dangerous than any weapon," — said the journalist of the edition of The Next Web Tristan green after testing AI algorithms.

In the future, the technology of text generation can be used to create books based on existing works of famous authors. Release date full version of the GPT-2 is not declared.


Source: http://4pda.ru/2019/08/28/360858/


RELATED MATERIALS: Science and Society
Возрастное ограничение