By Tunicia Phillips
There’s a new trendy hashtag on twitter and it’s flipping the lid on the ever-evolving war on truth. #DeepFake is the word tech nerds are using to describe a new way in which artificial intelligence is shaping the world of fake news.
‘Deep fake’ is the use of artificial intelligence (AI) to manipulate videos in order to make a subject say whatever the designer wants them to say. The technology was largely designed to make the film industry’s lives easier from an editing perspective. Today, even the technology’s developers say they are not naïve enough to believe that ‘deep fake’ or ‘deep video portraits’ will not be used by the more sinister fake information peddlers online.
The pornography industry has jumped onto the bandwagon using celebrity faces in porn videos (without their consent) but it is highly unlikely that the political underworld of disinformation will give the use of AI a pass. What does this mean for ordinary content consumers who still believe that ‘seeing is believing’?
It means that South Africa’s 30.8 million internet users and 16 million Facebook subscribers are all at even more risk of falling victim to disinformation. Added to this are other innovations like Adobe’s VoCo which allows you to make anyone say anything just by typing and using their voice. So while an online user could potentially use different frames from a fake video to reverse image search on google to verify its authenticity, it will be much harder to determine whether President Cyril Ramaphosa, for example, really asked his supporters to grab land.
Artificial Intelligence: Cause and the Cure
Artificial Intelligence is unavoidable in the age of big data, which is catapulting the world into the Fourth Industrial Revolution. At the same time, social media users are increasingly failing to distinguish between fact and fable and more tech companies are investing in solutions to curb fake information from spreading like wildfire across the pages of billions of online users.
While Artificial Intelligence is giving fake news a boost in innovation, AI and big data can also help stop the spread of disinformation. WhatsApp made headlines recently when it introduced a new update that prevents a message from being shared more than five times while Microsoft has introduced a built-in add-on to its Edge mobile devices that lets people know when information or news online is fake.
Facebook has committed to installing transparency tools for election advertising after a global uproar over Cambridge Analytica’s leaked data scandal last year in which 87 million Facebook subscribers’ personal data was harvested through a quiz application and sold to the British political consulting firm which specialises in data mining, data analytics and strategic election communications. 60 000 of the leaked data was from Facebook subscribers in South Africa. The firm was used for both US President Donald Trump’s election campaign as well as Britain’s Brexit campaign.
In the run-up to the 2016 US election, analysts found that fake news gained more engagement and shares on social media than credible news outlets. Buzzfeed’s founding Editor, Craig Silverman, told PBS News Hour that it was shocking that a four-month-old fake news site received more shares and engagements than credible news outlets like the Washington Post.
Ways to prevent the spread of fake news
In their paper titled, “Who falls for fake news? The roles of bullshit receptivity, overclaiming, familiarity, and analytic thinking” – Gordon Pennycook and David Rand state that the “belief in fake news has similar cognitive properties to other forms of bullshit receptivity, and reinforce the important role that analytic thinking plays in the recognition of misinformation”.
Researchers use the ‘bullshit receptivity scale’ to determine people’s inclination to believe conspiracy theories. The paper goes on to say that, “inaccurate beliefs pose a threat to democracy and fake news represents a particularly egregious and direct avenue by which inaccurate beliefs have been propagated via social media”.
The unfiltered nature of information shared on social media means that disinformation/ misinformation spreads fast, and the faster it spreads the more credible it appears. This means that a tremendous responsibility now rests with ordinary people to verify information before sharing it. The viral culture online is forcing users to become citizen journalists. If we cannot discern what the motive behind a fake post is, we can’t possibly predict the potentially devastating outcome of its spread.
Disinformation is the deliberate attempt to sway public opinion in a certain way to achieve the desired outcome. It is tantamount to information warfare. It is an age-old tool and has historically, contributed to the legitimisation of human rights abuses. The Holocaust is one such example of the power of propaganda.
Today, fake news is disinformation that has found a credible space in the form of ‘news’, while misinformation is information that is untrue but the intent was not to deceive. It is the equivalent to a factual error. Here’s what you can do to verify information and limit your part in the spread of fake news ahead of this year’s national election:
- For images, use googles reverse image search option which will locate the origin of a picture.
- For videos, screen grab frames and use the same reverse image search
- For information, check, check, check. As a rule of thumb, if you can’t trace the source of the information; or you can’t determine whether a profile or source is credible, don’t share it. This can be hard in the age of bots and websites specifically designed for information wars but, as a user, it is better to be wary of all information that does not originate from a popular and credible source of information.
Identifying fake news websites
According to tech site MemeBurn, there are five prominent fake news sites in South Africa:
– T1mesLive, the website name looks like the actual, credible news outlet but it has become notorious for announcing fake celebrity deaths.
African New Updates, like T1mesLive, this site masquerades as satire but many of their posts are click-bait headlines that are indeed fake news.
iMzansi, another site with a fine print disclaimer that admits to being fiction.
Live Monitor, according to Memeburn, gets fake news from other fake news sites.