Can AI be trusted to filter out fake news? How to navigate the internet with false content

American researchers have created an artificial intelligence chatbot to filter out fake news and stop the spread of harmful misinformation; can AI be trusted to detect fake news or are the people behind the chatbot pushing their agendas?

Can AI be trusted to filter out fake news?

Misinformation is one of the biggest causes of discrimination, fear, and judgement in the modern world; despite our access to the so-called “limitless knowledge,” we may be even further away from truth than in the past, threatened by the increasing spread of digital misinformation.

A recent case of former US President Donald Trump´s claims that Haitian immigrants in Springfield, USA, are eating the pets of locals, have led to more than 30 bomb threats in the area. The hatred for foreigners and xenophobia have been supported by internet sources as media outlets have used Trump´s claims to gain traffic to their websites and make a profit, despite the lack of facts and common sense in their content.

AI chatbot, DebunkBot, developed by researchers at MIT and Cornell University was launched amid the misinformation mania, spread by the former President and his supporters. The bot has been trained to filter out fake news by first appreciating the question and understanding its context and then using proven facts to provide truthful information without taking opinionated sites.

When asked about the effects of immigration, phrased as “Does immigration increase crime?” DebunkBot, to the surprise of many internet users, said “No, on the contrary.” “Multiple studies have consistently shown that immigrants are less likely to commit crimes than native-born citizens…” stated the bot. “FBI crime data indicate that areas with a higher number of immigrants often experience a decline or stabilization in crime rates, rather than an increase.”

Thomas Costello, a psychology professor at American University in Washington and MIT researcher emphasised that engaging with the bot can help filter out fake news more than people, as it has been trained to stay free from opinion and the “socio-psychological processes,” which have impacted people´s views. He noted that up to 50 per cent of the US population believed in Trump´s claims but after talking to the DebunkBot, a 20 per cent reduction in the participants´ belief was recorded.

“We found that the chatbot was making people less conspirational in general, and also increasing their intentions to ignore, block social media accounts, stop sharing conspiracies, or avoid dialogue with people who espouse these theories. It works,” added David Rand, another researcher at MIT. 

Google DeepMind researchers, however, stated in the paper, The Ethics of Advanced AI Assistants that even AI bots are under the threat of being influenced; “AI assistants may provide ideologically based or otherwise partial information to users in attempting to align to user expectations. In doing so, AI assistants may reinforce specific ideologies and biases and compromise healthy political debate.”

Take control and filter out fake news on your own 

When navigating the web, it is essential to stay weary of misinformation, take everything with a “pinch of salt,” and conduct further research to filter out fake news. First, when reading news, consider the source; news about the war in Ukraine, for instance, differ drastically on Russian and Ukrainian websites, each pushing their government´s agenda. Consider why the author is publishing certain content and what platform they are using to do so.

A description of a certain product, for instance, will certainly differ on the product´s original website, where it is promoted for sale, and in reviews from real people who have used the4 product. Moreover, read beyond the headline and the opening; go through the text analysing the sources and quotes used, keeping in mind that quotes especially, are often taken out of context and should be researched.

Where do the facts come from? Consider the supporting sources used by the author and the credibility of their statements; is it possible that the author is promoting their own beliefs? Read other sources and compare. If facts differ, there is clearly an issue. Also, check the date of the content to make sure it isn´t outdated and is relevant to today. Lastly, make sure you´re not reading a prank or a cruel joke by checking the source thoroughly and gathering evidence of their past credibility. 

Safe researching; Read more technology news.