` `

The Contradictory Roles of AI Tools in the Misinformation Battle

Khadija Boufous Khadija Boufous
1st August 2022
The Contradictory Roles of AI Tools in the Misinformation Battle
Bots can help propagate misinformation (Getty).

On many occasions, social media users deny and discredit the unlimited abilities of some Artificial Intelligence tools. Others doubt the existence of such inventions. While AI can help combat misinformation, it can also propagate false and misleading stories online. Misbar has outlined how automated bots spread misinformation. Misbar's blog explains that these online bots, also known as Artificial Intelligence tools, can spread misinformation by tweeting false information or using misleading content when commenting or responding to real users.

What Is Artificial Intelligence?

According to Professor Afafe Annich of the Higher Institute of Journalism and Communication ISIC in Rabat, Morocco, artificial intelligence, is a simulation of human behavior on multiple levels. Professor Annich explained to Misbar that artificial intelligence is based on learning and acquiring knowledge that will later be applied in various situations. 

"Artificial Intelligence allows replicating humans’ capabilities and performing tasks requiring human intelligence," she explained.

Professor Annich explained to Misbar that while the AI process may appear simple to the general public, it may require a number of complex stages and software to perform a simple task using an AI tool. She also mentioned that large datasets can help in the machine learning process.

Artificial Intelligence Can Spread Misinformation

The media's reliance on artificial intelligence can have unexpected consequences. Essentially, AI uses technology to mimic the process of creating content and sharing it on social media. Fake photos of people who never existed can be produced, and videos simulating real-life movement and speech can be generated.

As the production of audio deepfakes also became affordable recently, using Artificial Intelligence tools to simulate a real-life voice tone, verifying the content's authenticity became ever more difficult.

Professor Annich asserted that Artificial Intelligence uses Emotional Analysis techniques in creating content for social media platforms in order to make it similar to the work of real users and fill in the gaps that distinguish fake content from accurate content.

“Twitter bots have contributed to misleading the public opinion and changing its conceptions regarding issues like elections. This process relies on artificial intelligence and passes through multiple stages, starting with creating fake accounts using digitally altered profile pictures and perhaps producing fake audio clips and digitally generated videos. The process eventually leads to spreading false information that affects public opinion and directs the global attention for certain goals,” the expert told Misbar.

Artificial Intelligence Can Defeat Misinformation

Since artificial intelligence produces a lot of misleading content, debunking the false content requires a reverse process. For the expert, this task concentrates on identifying the altered or fabricated photos or identifying any information that will help debunk the content. “We can also reverse search texts using AI to find whether they were generated or written by normal social media users,” Professor Annich noted.

Despite the fact that many experts question algorithm improvement and updating, particularly in terms of reducing the duration and sources of process completion, Professor Annich told Misbar that the scientific research area is rich in terms of algorithms that can detect fake content.

“There are two groups of algorithms. The first group focuses on discovering the content and ensuring its authenticity, whether it is texts, images, audio clips, or videos. This first group relies on deep learning and machine learning in addition to large databases that can help spot the false content,” the expert noted. “These algorithms also use the Natural Language Processing NLP to analyze, synthesize natural language and speech, and confirm whether some content is generated,” she added.

The second algorithm type, according to Professor Annich, uses News Propagation tools to track the spread of false content. To detect "fake news," this group uses links between fake content on social media and user interactions. To achieve deeper insights, the Graph AI technique also employs machine learning and focuses on the relationships between variables.

“The Graphs are networks that provide a set of links expressing a connection between the content and different platforms and groups in addition to the interaction occurring in all these elements,” the expert told Misbar.

Automated bots, rather than humans, are the best way to use AI to combat misinformation, despite its potential for conflicting results. According to the expert, raising awareness is crucial for social media users. While fact-checkers cannot keep up with the continuous spread of information, media organizations should allocate additional resources to investigate and verify the accuracy of the stories they publish.

Misbar’s Sources:

Pulse Ghana

The Sun

Harvard Magazine


Read More

How Do Online Bots Propagate Fake News?

A New AI Tool to Counter Misinformation on Wikipedia