By Tiana Cline, Biz-tech Journalist
Fake news might be an old problem but there are new-world solutions to tackle it. Synthesis Technologies, along with some well-known writers and journalists, took a deep dive into the state of media to discuss the fake news epidemic and Artificial Intelligence (AI) as the potential remedy.
When it comes to fake fashion or forged artworks, it’s relatively easy to work out whether an item is authentic thanks to technology. Powered by AI, a company called Entrupy has collected thousands of images of physical goods from around the world, creating a rich and diverse unmatched data-set. Every item that’s authenticated helps the algorithms learn and improve, creating a smarter solution that adapts to the changing world. Last year, Amazon released Project Zero. Powered by Amazon’s machine learning (ML) technology, Project Zero continuously (and autonomously) scans their stores to proactively remove suspected counterfeits.
But was does fake fashion or counterfeit products have to do with fake news? And is there a technology solve to the global problem of misinformation? Dr Ted Ross, who is the head of the World Health Organisation (WHO), has said what we are facing with Covid-19 is, in fact, a double, triple pandemic.
“Fake news has become such problem in our lives, in our democracies, in our world, that we’ve moved towards different names to call it by. Now it’s called misinformation, that which is spread without necessarily maligned intention, and disinformation – that’s the really dangerous one,” explains Ferial Haffajee, Daily Maverick’s Associate Editor. “An example is your auntie sending every bit of Covid-19 information she receives over WhatsApp. Yes, some of it is benign – we all know garlic and ginger is really good for us – but often the information is dangerous. It’s an information pandemic. And it goes from your aunties to the really dangerous stuff like the 2016 US election, where technology was used or misused.”
There’s no doubt that misinformation is potentially dangerous. The BBC reported that conspiracy theories, rumours and fake news have killed more people than the coronavirus pandemic. From drinking cleaning products to cow urine, Covid-19 is an infodemic with serious implications. But is there always ill intent around the spread of fake news or are people ignorant or just distracted?
“Humans are so important in this is because it’s about developing relationships and a lot of investigative journalism may be about crunching data – and that’s where AI comes in – but it’s also a about the nuance and the personal feel as well,” says Mandy Wiener, author and 702 Radio host. “A lot of investigative journalism is based on trust. It’s about the relationships, what people are saying to you, and being able to kind of roll that information over and look at it from every angle, and try and really assess for yourself whether or not it is fake or not.”
People just don’t necessarily understand that not everything they read is true yet it is important to be discerning. Social media has only added to this complexity – we can no longer rely on the ‘written word’ and curated news feeds show one side of the story.“There’s a few issues with social media that create and manifest this problem. So the one is the sheer volume of news that’s being produced. And because of social media, everybody is a citizen journalist, and there’s just so many sources of information,” adds Wiener.
Archana Arakkal is a Machine Learning Engineer at Synthesis Software Technologies who has done extensive research into fake news, AI and social media. Along with her research group, she’s been creating patterns of how fake news is spread:
“What we found is that it is very similar to how a pandemic or an outbreak actually occurs within a space and fake news travels a lot faster than real news. The fact that fake news is unfiltered, and the fact that fake news isn’t checked with credible sources, is the reason why it spreads so fast. So if something is spreading really fast, the likelihood of actually being fake news is quite high,” explains Arakkal. “With real news, as soon as all the credibility checks come into place, it takes a lot longer to propagate.”
Fake news is a disease in its own sense. A recent study showed that out of the 98,000 tweets which are sent out on a daily basis, close to 70 percent of those tweets are, in fact, fake news. In 2010 Malcolm Gladwell wrote that revolution will not be tweeted.
“AI may be good for spotting a fake bag, but I do not think it is the way that we will deal with misinformation and disinformation. Facebook’s data scientist, a man called Eduardo Ariño de la Rubia, said: AI fundamentally cannot tell us what’s true or false. This is a skill much better suited to humans. Public pressure has forced companies like Facebook and Twitter to be better…” adds Haffajee.
“You can’t ignore the fact that social media exists, you can’t put your head in the ground and just pretend that that doesn’t happen. It is definitely an echo chamber – it is not the real world,” says Wiener.
Another issue is that often, AI is used to create and spread misinformation. Bots replying to social media posts, for example, are often phishing attacks. A study done by the University of Chicago showed how AI bots could easily write restaurant or hotel reviews using data sets found online. What’s more is that those who read reviews generated by AI next to a real reviews couldn’t tell if AI or a person wrote the review. While false restaurant reviews seem relatively harmless, AI and ML can be used to distort our perceptions of reality, especially in this era of fake news.
“Are we using AI to fight AI? Is AI the solution? There are many tools out there to synthesise video and sound and it is incredible realistic… so what happens when you can no longer trust what you see or hear? A lot of researchers are moving onto next-gen technologies like blockchain as a solution to stop fake news,” explains business-technology journalist Tiana Cline. “You could use a blockchain ledger to have a level of transparency and accuracy in order to provide news content that cannot be altered.”
Technology has really matured over the past years but we need to ask if there a way we could use AI, ML and other types of next-gen technologies to successfully tackle the fake news infodemic when a lot of the consumption and distribution of misinformation lies in human error. Luckily, there are many algorithms and different technology stacks that are able to support big data and facilitate these new ways of working.
“We need to leverage on this technology to ensure that the correct information is being propagated because it’s a two-man approach. We need to work with the tech in order to utilise it to a good extent. It does eventually merge together but there needs to be a sense of education around what tech can and can’t do… so that we can utilise it in the most effective way possible,” ends Arakkal.
Did you miss our fake news webinar? Catch it on online here: https://www.youtube.com/watch?v=irHP3znqwq8
