It seems barely yesterday that most of us had never come across the term ‘fake news’. Now, it crops up everywhere – Donald Trump frequently blames it for negative stories about him, while Germany has taken legislative action to tackle it. One of the defining features of the rise of fake news is the role that technology has played. Here, we explore how AI research is being abused to create fake news, as well as how AI can be employed against it.
What part has technology played in the rise of fake news?
The famous saying tells us that a lie can get halfway around the world before the truth has got its boots on – though fittingly, no one can agree who said it. Half-truths and outright falsehoods being reported as fact, whether unknowingly or maliciously, is hardly a recent phenomenon, but in the last few years a new, distinct trend in factual misreporting has arisen – and it’s been fuelled by technology.
For most, the first time they heard the words fake news was during the 2016 US Presidential election campaign, when bogus stories – ranging from non-existent endorsements to disturbing allegations – began circulating on social media. Our increasing use of social media as a source for news and information, coupled with the algorithms that determine the content we see on our social media feeds, has been central to the spread of fake news. The internet may have given us access to the greatest pool of human knowledge in history, but this freedom comes with a darker side – fake stories are easier to create, often hard to spot, and even harder to police.
Recently, an AI developed at the University of Washington was able to analyse footage of President Obama and then synthesise an eerily accurate virtual facsimile. Similar technology for mimicking voices is becoming increasingly advanced, as is an AI that can learn from real locations to create believable virtual environments. People, their voices and their surroundings can now be recreated with a worrying degree of accuracy – the potential for this technology to be combined and abused by the creators of fake news is self-evident.
How can AI be used to tackle fake news?
AI’s potential use in stoking the fires of fake news comes with a flipside – as the technology that creates and spreads these stories becomes more advanced, so does the technology capable of fighting back.
In particular, many see AI as a tool to perform the fact-checking of fake news. According to Catherine Lu from the fraud detection company DataVisor, AI could be used to analyse the content of viral web stories, then compare its contents to other reputable sources and recognised facts. Researchers at one university are already working on a machine-learning approach to the issue, that would give each article it analysed a score based on the likelihood of the story being false. Although these projects are still in their research and development phases, their potential is significant – theoretically they would become more accurate and effective with every false story they identified.
If the misuse of AI does lead to ever more complex and believable examples of fake news, we can be hopeful that the challenge will be met in a similar fashion – with equally advanced AI applied for the right reasons. Our own founder Tej Kohli is a firm believer in the potential for technology to be used for positive social impact, and the responsible use of AI in the service of the truth would be of benefit to us all.