A new AI is Using News Outlets to Produce Fake Article and it is quite Concerning!

Over the last few years, the rise of fake news and information has been off the charts. To tackle this issue, various measures have been implemented. Most recently, the Allen Institute for Artificial Intelligence developed Grover.

Grover is basically a neural network that can produce fake articles by imitating the writing style of actual journalists. The motivation behind its creation is to identify fake news, and generate fake stories as well.

The Institute issued a statement confirming the above motivation. It was claimed that detecting neural fake news requires a model capable of generating fakes. Grover will be aware of its own characteristics, behaviors and twists, as well as those of other similar AI models trained using similar data i.e. publicly available news.

Fake news is generated and spread both by humans as well as AI systems. Although a reader can easily verify the authenticity of an article, most people are simply going to believe what they read.

The creation of OpenAI’s GPT-2 is yet another indication how the task of generating articles in the future will be taken over by AI systems, and with such accuracy that people will find it hard to differentiate between AI-generated articles and the ones created by humans.

Coming back to Grover, it was tested in various ways. In one case, it was found out that by simply providing the name of Author, headline and news outlet, a well-written yet inaccurate article could be produced. For readers with no knowledge of a particular subject, even an inaccurate article would be considered genuine.

Another example article was also generated where President Barack Obama mentioned that it was time to start supporting Donald J Trump. The material generated by Grover in this case was shockingly identical to Obama’s op-ed style in the New York Times.


Grover is quite simple to use and that’s concerning. With the quality of content it is able to come up now, it can be said that within a decade or two, it will generate articles in such a way that no one will be able to tell whether the content is produced by real writers or by algorithms.

Fortunately, Grover’s purpose is to battle fake news by identifying the AI-written articles. It does so with an impressive 92-percent accuracy.

For now, you can test Grover as well by visiting the official website of the Allen Institute for Artificial Intelligence.

This terrifying Artificial Intelligence creates fake blogs from any news site
Photo: Getty Images/Science Photo Library RF

Read next: Teachers can use Machine Learning to Determine if their Students are Understanding the Taught Concepts!

No comments:

Post a Comment