Elon Musk is known, among other things, for harshly criticizing artificial intelligence (AI) software. He points out its negative possibilities. Interestingly, according to what MIT Technology Review reported, one of his companies has just created an AI that can automatically generate fake news texts.
Needless to say just how destructive such software can be in the world. As people today increasingly believe in fake publications and dubious sources. But it is interesting to note that if such a program falls into the wrong hands, all the work done by social networks like Facebook and Twitter to reduce the spread of fake news on their platforms can go down.
It’s one thing to use machines to compete with humans doing manual labor. Another totally different one is to fight intelligent machines capable of creating credible but false texts.
The software was created by OpenAI by Elon Musk. But the company is considering not sharing the code with the world for obvious reasons. MIT Technology Review says that the company may only publish a simplified version for scientific purpose. It would not have the ability to actually generate good texts for false news.
They created the original versions of the software for translation and other text processing. The creators, however, soon realized that the program would be capable of something different.
Most of the pieces have serious problems of cohesion. This prevents the reader from understanding.
Fortunately, even the full version of the program is still unable to produce mass speech materials. According to its creators, most of the pieces have serious problems of cohesion. This prevents the reader can understand what the software really meant.
The program also does a lot of direct plagiarism. So it would be easy to detect when such text was published on the web. Even so, time and time, AI can create really elaborate texts. According to OpenAI, to find out if the information in these pieces is false or not, it takes a very hard work of fact checking.
In any case, OpenAI’s decision not to distribute the software is only an ethical action by one company. Other several or even government agencies may be doing research on the same line right now. So it may be that in two or three years the internet is even more infested with false news than it is today.