Deepfakes are pieces of synthetic media, which in recent years have become increasingly common on websites such as YouTube and Reddit. They are completely fake videos, manipulated to depict people in a completely false manner, and involve replacing an existing image or video with someone else’s likeness. Despite bans on major social media platforms such as Facebook, they remain easy to produce, with programmes to construct them found all over the internet.
On the surface, many deepfakes are comical and appear rather harmless. For instance, videos of Keanu Reeves’ likeness being inserted into the movie Forrest Gump have widely circulated. Garnering attention because of their humorous novelty, it makes us wonder why they are such a problem - surely, they are not intended to cause harm or offence?
As modern technology has rapidly improved over the years, there is no surprise that deepfakes are now becoming more sinister. Now, many deepfakes can pass off as remarkably real, often being used to shock and upset viewers.
The viral deepfake video of Barack Obama describing Donald Trump using a derogatory term just demonstrates how believable these videos have become. If I didn’t tell you that this audio recording of Bernie Sanders reciting the Communist Manifesto was actually an algorithm pretending to be him, would you have realised it? Would you have believed it if you saw the clip shared on your Facebook timeline or Twitter feed? I certainly would have.
The fact that many of these videos now feature politicians making alarming remarks should be worrying for us all. According to a Pew Research poll, 63% of Americans say ‘made-up or altered videos and images create a great deal of confusion about the facts of current issues and events’. With a deeply disturbing statistic in the backdrop, this remarkably high figure is a cause for concern. Yet just because the public is aware of deepfakes doesn’t mean that they will spot them every time. According to the same poll, 61% say it is ‘too much to ask of the average American to be able to recognise altered videos’. This suggests that more needs to be done by social media platforms and law enforcement officials to root out deliberately misleading content online.
Politicians and public figures are obvious targets of deepfake content because they are in positions of authority. People producing deepfakes know the high viral power these videos hold. The chances of popularity could increase as it is more shocking to hear outlandish remarks from politicians. Why? Elected officials’ statements have weight to them - sometimes the political consequences can be severe. Splicing up and editing videos for electoral advantage is nothing new in politics, but not to the extent that could soon be possible.
For example, during the Democratic primaries, Pete Buttigieg’s social media team increased the level of applause he was receiving at his town hall meetings. This was done in order to make him seem like his speeches were more appreciated than they actually were. Just last year, a video was spread online that seemingly showed House Speaker Nancy Pelosi slurring her words as she spoke. This turned out to be a digitally doctored video – the first major sign that doctored video content could be shared on social media by individuals with partisan motives.
Deepfakes were not an issue during the 2016 US election campaign, but the technology used to produce them has advanced rapidly in the past few years- to the point that they could be feasibly used as a method to spread disinformation.
Can you imagine the damage a nation-state intent on influencing a US election could do with a carefully planned digital attack? By producing high-quality deepfakes and strategically releasing them in a manner aimed at maximising their impact, the results of a targeted deepfake attack by an adversary could be disastrous. It could have the ability to sow the seeds of dissent within society, undermining our democratic values in the process.
According to a 2019 Gallup poll, the public’s trust in media is at a measly 41%. These figures would plummet even further if convincing deepfakes were in circulation.
People simply would not know what to believe.
We should all anticipate that foreign adversaries intent on interfering in our political processes will not act in the same way every time. Over time, they evolve their tactics, learning what they did in the past to try new things. For instance, the recent hack of Twitter, whereby hackers posted content on notable Twitter accounts, Bill Gates and Barack Obama among others, showed us that social media platforms are not fool-proof. They are easily vulnerable to manipulation by those with the high technological capabilities intending to cause harm.
Ultimately, it is increasingly clear that deepfakes are a problem that urgently needs resolving by those holding political power. It is commendable that the Pentagon has deemed deepfakes an issue of national security, but more needs to be done. to limit their spread.
Media outlets must be alert and wary of the sources and type of information they receive. Social media organisations have to be more robust in taking down content that violates their terms of service. Members of the public need to be more reluctant in sharing content that encourages sensationalism and outrage.
It sadly seems that deepfakes are going to be the next weapon in the attack against truth, and until these problems are addressed, our democracy will pay the price.