Why Deepfakes In Elections Are Containable, For Now
Deepfake technology and election manipulation are often cited together. And that’s not a coincidence, because politics are all about convincing people you’re the optimal candidate to lead the country. And deepfaked content is there to either reinforce or destroy that public confidence.
With the American elections of 2020 (between Donald Trump and Joe Biden) around the corner, the role of deepfakes on social media should rise, warn experts. The divide between the Republican and Democratic candidates has never been bigger, and the trust in televised media has never been lower.
It seems to be the perfect storm for deepfake tech and Artificial Intelligence (AI) to shine. Using these new technological tools to further a certain side of the political coin is probably more rewarding now than ever, because the regulatory framework to stop these tools from manipulating election results is pretty much non-existent.
The Impact Of Deepfake Tech On Politics
It’s therefore extremely interesting to take a look at the impact of deepfakes in the world of politics, especially for an election period in (probably) the most important global election in recent history. And it’s not just the elections. Other politically motivated groups could start using deepfake tech to further their agenda’s too. Think about politically motivated activists such as Antifa, Qanon and the BLM protestors, among others.
Interestingly enough, it seems like the web of confusion around ‘real news’ and ‘fake news’ makes the threat of deepfake videos and images relatively mild. Thus far, the manipulation on a massive scale has yet to occur. The political candidates don’t seem to have to use deepfake videos to alter the political bias, but can instead choose to pick more primitive information manipulation strategies. Let’s explore why this is the case, starting with the essential role social media plays in all of this.
Political Manipulation On Social Media: Deepfakes Or Fake Messages?
Disinformation on the internet moves fast, and there are plenty of malicious actors that try to take advantage of this. At the same time, containment or correction of such misinformation campaigns is slow, ineffective, and often pointless to begin with.
Because how do you even fight against a huge army of internet trolls?
Political campaigns have more important things on their mind than constantly scanning the internet at all times for false information about their candidates. Only when the avalanche has picked up so much snow it can’t be ignored anymore, a counter-statement would need to be made.But at that point, most of the damage has already been done. The question is: does it matter? The echo chambers on the internet have never been more prevalent.
The Digital Poltical Echo Chamber
People on Twitter, Facebook and Instagram often reside in their own little bubble, an echo chamber with a particular confirmation bias. Whether you’re in a Qanon Facebook group, or follow a lot of Antifa Twitter accounts: the effects are the same. It doesn’t matter of the source material is manipulated. You simply hear and read the things that will confirm what you already believe.
At least, that’s the sad reality for the bulk of the American public.
We cannot ignore the fact that the attention span for the average social media consumer isn’t that big. The majority of people will simply scoff and say something like: “Ha, I knew it!”. Deepfakes or other types of fake messages wouldn’t put a very big dent in that at all. Even though they make it harder to differentiate between what’s real or fake.
Algorithms Versus Digital Signature Standards
A very big factor in this growing confirmation bias is the almighty ‘social media algorithm’. Users on social media platforms will only get offered content that is similar to the one that is already being consumed. Large scale internet manipulation groups exploit this digital reality to the fullest. Controversy sells and generates clicks, after all.
Is there anything defenders of the truth can do, apart from fighting the avalanche of lies?
Yes, and one of the solutions is perhaps elegant and simple. Take the Content Authority Initiative (CAI), launched by Adobe, The New York Times Company and Twitter. The aim of the CAI is to develop an industry standard for content attribution.
Placing a digital signature within the video, image or other media can easily prove if something is real or fake. And it will silence the majority of people claiming that a piece of falsified media is truthful. To further strengthen such digital signatures, one could incorporate it into another growing digital industry, the blockchain world. There are plenty of blockchain-related projects that attempt to create neutral, decentralized immutable systems for information verification.
But not every large social media network will go as far as creating a digital signature standard. Places like Facebook are still the wild west of Qanon and Antifa groups, with Facebook only recently taking a stance against the extreme sides of these types of movements.
Incorporated in this push against misinformation, Facebook formally made deepfake videos illegal in the run-up to the 2020 US Presidential elections. A nice gesture, but probably just an empty statement. The policy didn’t include so-called shallowfakes, which reside within the grey area between deepfake videos and real videos.
Furthermore, the entire premise of deepfake content is to be completely indistinguishable from reality. So one could question how the army of Facebook moderators would push back against an invisible enemy. AI-manipulated content will spread much faster than the ‘experts’ can detect them already. Enforcing it in practice would be a monstrous (and probably impossible) task.
Beyond Social Media: How Deepfakes Can Influence Your Vote
We live in strange times. The President of the United States has openly questioned the legitimacy of the American democratic system by opening the attack on mail-in ballots. Let’s ignore the possibility that these claims might be false, just for the sake of argument.
With the rise of deepfakes and related AI technologies, societal truths are clearly under attack. Right now, 70% of American citizens actively doubt the government because of fake news. And it doesn’t even matter if you’re in camp Trump or camp Biden. These issues are beyond anyone’s political color: our collective perception of reality is under attack.
The threat seems to be a bigger concern than many other major political issues. Even terrorism, immigration and racism issues were smaller collective problems, a Pew Research Center survey found.
Deepfake videos and images are just a small part of the larger problem of ‘information manipulation’. But they will likely play a bigger role in these elections, especially since we know that Russians are likely to meddle and manipulate with large-scale internet misinformation attacks. With deepfakes widely available and on the rise, the average voter will inevitably be influenced at some point.
Deepfakes are a means to an end – voter manipulation – which will likely be a contributing factor to a larger set of manipulative factors. It’s true that deepfake video creators will always be one step ahead of those trying to stop them, therefore, the critical thinking skills of the voter must be the last line of defense. It is now more important than ever to fight for the truth.
Fighting The Fake News: Critical Thinking Without Echo Chambers
Deepfakes and fake news go hand in hand. But the existence of deepfake content on the internet can disrupt any election process, without the deepfake ever coming ‘to the surface’.
It’s unlikely that deepfakes will screw up the 2020 elections. Kathryn Harrison, CEO of DeepTrust Alliance, stated it clearly on CNET: “If you were to ask me what the key risk in the 2020 elections is, I would say it’s not deepfakes”. The reason is simple: information manipulation is much bigger than deepfakes alone.
And let’s not forget the fact that the divide between Trump’s and Biden’s base might simply be too big already. Trump’s base claims that fake news media exists. Biden’s base claims that the White House is run by pathological liars. Whatever side you take in this political finger-pointing game, the core problem persists:
Misinformation campaigns are here, and they are here to stay.
So regardless of the echo chamber of right-wing or left-wing news sources you’re in, people need to start equipping themselves with a new set of critical thinking skills. Perhaps we need to start with our youngest generation and put the education system to work. Is someone willing to teach the American children some Deepfake 101 classes?
The Liar’s Dividend
There’s the education solution, we can put digital signatures in our content, and we can hunt down troll farms. It’s nice to put up a fight against deepfakes, but it won’t win the war. There is still this underlying, looming threat, one that’s especially easy to use in the political world.
Deepfake experts refer to this threat as ‘the Liar’s Dividend’.
It means that the existence of deepfake content gives more credibility to deniers of the truth. Lies become truths. Truths become uncertainties. The largest victim of it all will be reality itself, and the fabric that holds societies together will weaken.
To emphasize this point, we probably need an example of a not-so-impossible scenario:
Let’s say you’re the President of the United States, and you get caught on videotape sleeping with women that aren’t exactly the FLOTUS. The President is a smart man. He does what politicians do best. An excuse is made up on the spot to defend their political position.
He can do it in this case, because he’s using the Liar’s Dividend. The videotape could easily be a deepfake video, right? The enemies of the state must have created this smear campaign against the President to weaken his political position!
It doesn’t matter if the President is guilty or innocent in this case. The possibility of him being guilty can simply never be determined because deepfakes exist. It’s a solid excuse for everything a President will ever say or do. Because the public will never find out. They also know fake media exists. They know deepfakes are a threat to democracy.
Deepfakes: A Symptom of a Larger Societal Problem
The world is now more interconnected than ever, but the primitive misinformation techniques still seem to be the most popular for the 2020 elections. Deepfake videos or images probably exist out there too and could pose a minor problem, but there is currently no evidence of large-scale use of the technology as a manipulation tool.
The type of misinformation tool that’s being used is actually not that relevant. It’s the interconnectedness and polarization in society that is the crux of the problem. Most people take their information from Google, YouTube, Facebook and Twitter these days. People are in chat groups on Whatsapp or Telegram, each within their own little political bubble.
The companies that draw you in, with their sophisticated algorithms to keep you coming back, are at the core of the problem. It’s the Facebooks of this world that have made it their business model to show you as many relevant advertisements as possible. The more information you like, share or comment on using on their platform, the bigger the eventual profit for the platform.
The algorithms are designed to show you a controversial standpoint, because it creates engagement. Extreme stories are interesting. Emotions evoke interest in people. The more polarized the message, the more people want to engage with it. This is the complete opposite of showing nuanced news stories based on facts and political neutrality. Boring isn’t captivating.
The sober truth is drowned in a world of sensationalism and extremism. A two-party political system such as the American thrive well on it, but also get driven apart by this new way of information provision. Political parties can even target advertisements in a way that their messages only target the key swing states, or even easily manipulated people. Who needs deepfake videos when normal manipulation is easier than ever?
Primitive Manipulation Triumphs Deepfakes
For now, and likely also for the 2020 US Presidential elections between Trump and Biden, deepfakes only play a minor role in the eventual political outcome. That does not mean information manipulation isn’t a major factor of influence in the electoral process. In fact, it might be more important than ever!
The interconnected world allows for classic and somewhat primitive social manipulation methods to thrive and prosper. Who needs deepfaked videos, when you can simply send an army of trolls into a Facebook group? When you can put Qanon and Antifa groups up against each other with the click of a few buttons?
Deepfakes are just one of the many manipulation tools used in (social) media. They might be the precision long-distance cluster bomb, whereas now people still like to bonk each other’s heads in with sticks and stones.
When the social media-dominant internet landscape changes and algorithms no longer thrive on extremism and polarization, deepfakes might become more relevant again. It’s only when the small nuances make a big difference, when deepfake videos might turn out to be a more effective information manipulation tool. It seems like deepfakes are still somewhat of a contained problem in the wider electoral process. Let’s see what the next few elections might bring.