The OpenAI research group has demonstrated artificial intelligence (AI) that can compose authentic-looking fake news articles from a few fragments of information.
After being fed a few sentences of sample text, the software successfully generates a persuasive, but completely false, seven-paragraph news story.
The AI was trained to perform language modeling, or predicting the next word of a piece of text based on knowledge of all previous words.
OpenAI's Jeff Wu suggested the software could have beneficial applications, like helping creative writers generate ideas or dialogue, or hunting for bugs in software code.
New York University's Sam Bowman and OpenAI policy director Jack Clark agreed the AI's abilities are currently too inconsistent for it to be weaponized for spreading disinformation.
From Bloomberg
View Full Article
Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA
No entries found