Can Artificial Intelligence be used to detect fake news?

Ved Panse

We heavily rely on social media nowadays as a source of information. Around 27 billion texts get exchanged daily worldwide. Have we ever thought of how many messages we receive are fake?

The Stanford Graduate School of Education conducted two studies on 3,446 American high school students to analyse how they perceive unfamiliar news. In one of the studies, students examined a website that reported climate change is not that bad. 96% of the students failed to discover that the publisher was related to the fossil fuel industry and had ulterior motives. This study reveals that the structure, design, domain name, and other weak factors of a website easily influence students. This raises concerns about the safety of the modern world.

To understand this problem further, we need to identify the psychology behind creating such news. Consultant psychiatrist Dr Tejas Gayal mentioned that the people fabricate information with a motive to align others with their opinions, which could also be religious or state-sponsored.

The spread of fake news has led to furore several times. We are witness to unverified remedies for curing the COVID-19 circulated in India.

Can we use the current technology, especially Artificial Intelligence (AI), to tackle this issue? It is supposed to make computers think and work like human beings. Social media giants are already developing structures based on this technology.

Facebook, accused of being the source of fake news on the internet for a long time, has built an AI technology called SimSearchNet that detects manipulations done to a picture and helps them find a fake image with other tools utilized.

Similarly, Google has built a fact-check explorer. Typing in recently airing news will tell you if it is fake. A human team is assigned to go through the final verdict. So, it is more reliable but tedious and hard to maintain by a unit.

There is a strong argument against fake news censorship for being against the Freedom of Speech. In the Philippines, Senator Joel Villanueva has introduced a bill that allows up to five-year prison terms for those involved in publishing and spreading fake news. However, this bill has invited a lot of censure since punishments can be arbitrary. It is difficult to ascertain whether the restrictions imposed to stifle harmful fake news are proportionate to the objectives established.

Eventually, we must remember AI works on data provided to these computers just like parents teach their kids. Naturally, it will have its own biases based on the parent. Assuming that we come to some common grounds on this, these approaches might seem good enough to contain the spread of fake news.

Yet, a lot more questions remain for us to ponder

1. Can AI technology be fully trusted to help us avoid fake news?

2. What criteria do we feed to the AI to distinguish fake news?

3. How can we make this process more transparent?

4. Can AI comprehend false information in different human-spoken languages?

5. Can AI be used in corner cases (cases involving a variety of perspectives and speculations)?

Though there are mixed opinions, let us hope that this research yields some results and prevents the spread of fake news without infringing on our freedom for the betterment of humankind.

(The writer is student, Woodridge High School).