5 Ways In Which Deepfakes Are Used By People To Manipulate
Deepfakes are modern tools of manipulation. They have garnered widespread attention for their use in fake news, fraud, revenge pornographic materials, as well as political manipulation. And that’s just a drop in the bucket in the early stages of deepfake technology.
Where a tool can be used to further a person’s agenda, it will be used to the full extend.
It’s obvious there is a real threat here, one that is actively being responded to by both tech companies, media outlets, as well as governmental organizations. Despite the combined effort and massive resources of organizations around the globe, it will be extremely difficult to stop the technological advancement in its tracks.
This article aims to explore the five main ways in which deepfake videos and images are used to manipulate society on every level, from the abuse of global organizations to individual fraud or blackmail. But let’s start at the basics, and explore the concept of deepfake technology a little better. Knowing what we are dealing with is, after all, a key element in detecting the problems and potentially fight them in the future.
What Are Deepfakes?
A deepfake is a video, image or audio content piece in which a person is replaced with someone else’s likeness, thus creating new ‘fake’ synthetic media content. Deepfakes are automatically created by computers using machine learning and Artificial Intelligence (AI) algorithms. Special deepfake software is utilized to apply these algorithms to existing media content.
The technology used to create deepfakes is, at the time of writing, relatively new. The term first became popularized in newspapers and other news media outlets in early 2018, when the real surge of the deepfake ‘threat’ reporting started to emerge. As we roll into the 2020’s, this treat has only increased on an almost exponential scale.
Some media channels characterize deepfakes as “the 21st century’s answer to photoshopping”. But it is way more than that. The AI tech and algorithms allow malicious actors to further their agenda on scales which have never been seen before. Allow us to describe some scenarios to showcase why it is important that manipulation with deepfake software needs to be addressed right now. Because all it really takes is one single person with a bad idea and some willpower to make another person’s life as miserable as possible.
All It Takes Is A Kid With A Laptop And Some Wi-Fi
A clever whizzkid with a simple laptop and an internet connection can, in theory, use the technology to alter an electoral result, or become a millionaire through fraud. When you combine a ‘decent’ malicious idea with a smart kid that has access to widely available open-source deepfake software, the bad stuff is bound to happen at some point in time.
It doesn’t even need to be large scale manipulation either, it can also be on a local or personal level. Or on a smaller and more perverted scale, the same whizzkid can fulfill its wildest erotic dreams through putting the head of his crush on, let’s say, Mia Khalifa’s body in one of her ‘adult entertainment videos’. And then he can decide to blackmail his crush and send the video to all her family members on Facebook, unless the poor girl pays random money.
It sounds like an episode of Black Mirror, but it’s not that far-fetched now that deepfakes are more easily accessible than ever. A personal revenge can turn ugly really quickly with a bit of deepfake software.
It just goes to show that deepfakes have more reach into personal and political spheres than most people can suspect. It will obviously require an ‘evil person’ with enough willpower and technological skills to pull these things off, but software is making it easier every day. And criminals or even hustlers looking for some quick money are not that difficult to find. So we can’t really say that the scenarios like the ones described above are unrealistic. Given enough time and technological development, they will inevitably start to happen.
How Deepfakes Are Used To Manipulate
The applications of deepfakes for manipulative purposes, to further the agenda of the few, are plentiful. We’ll get into the five main themes in brief in this article, but please note that the full spectrum of application is actually much broader than this.
Much of these themes are expected to have a large impact – as we are currently in the early stages of deepfake tech development, the larger impacts are still to be seen. As time progresses, it is expected to get a much better overview of the ways in which deepfakes can impact us all.
There will inevitably be manipulation themes that weren’t foreseen, so please remember that when consuming the information presented in this article. The first and perhaps also most important theme will be political in nature, as public opinion is very easily impacted in this part of global society. Apart from politics, we’ll also have a brief look at financial fraud, deepnudes and revenge pornography, identity fraud, and media manipulation through fake news or hoaxes.
1. Political Gain Or Political Manipulation
Politics is generally one of the central themes that are brought up as soon as deepfake technology is mentioned. And for good reason, as well. Especially election processes around the globe are at risk, now that political preferences can change faster than ever. The rise of social media has created a whole new world of risk for ‘the truth’, now that any ‘influencer’ can pretty much broadcast whatever they please.
One infamous case that is often brought up is the situation around an Indian politician, which used deepfake technology to win over new voters. Obviously with manipulation of the truth and targeting voters that could decide his electoral fate. The deepfake videos created by the President of the ruling Bharatiya Janata Party (BJP), mister Manoj Tiwari, reached an estimated 15 million people during his election process. Not a small effort, but perhaps the first serious case of deepfake manipulation by a politician running for office.
But it’s not a single instance. And not just in developing countries with ‘wobbly’ democratic processes. It happened in Italian politics as well. The ex-prime minister Matteo Renzi was used in a deepfake video creation that went viral on social media website Twitter. It caused an outrage and likely impacted the election process in the European country.
And the case of the American presidential elections is also a sensitive one. Most people will remember Russian meddling in the 2016 US presidential elections, which eventually elected Trump into office. A controversial outcome, since Russian cyberattacks and meddling had impacted public opinion in the USA significantly during the election period. All of this is well documented, albeit still very controversial for some groups of people.
The same could very well happen for the upcoming 2020 presidential race between current President Trump and his rival, the Democratic candidate and former vice-president under President Obama, Joe Biden. Maybe not created by the politicians themselves, but if meddling happened once, it is likely to happen again. Or at least an attempted media attack will be launched. More about the 2020 election-topic is explained very thoroughly in this article by the website ‘Towards Data Science’. We recommend you to give it a read.
2. Financial Fraud Through Phishing Or “Vishing”
But don’t think it’s just politics that can easily be manipulated. Deepfake technology provides a perfect tool for clever criminals and criminal organizations involved in financial fraud schemes as well. This usually comes in the form of audio deepfakes, often referred to as ‘vishing’ or ‘voice phishing’.
The term is obviously a play-on-words to the more commonly used term ‘phishing’. This is a fraudulent attempt to obtain sensitive information. For example obtaining credit card information, sensitive passwords, usernames, or PIN-codes. The criminal will disguise themselves as a trustworthy organization and send out an email.
Vishing is slightly different. Let’s say you have a scheduled appointment with a bank clerk on the phone about your mortgage. A cyber-criminal can use data leaks, hacks or inside information to obtain the details of the call (these are recorded quite regularly). Using voice-altering software, the criminal using vishing can then make the caller believe it’s the same person. Deepfake audio will often copy voice intonations and mannerisms perfectly, which makes them very difficult to distinguish from the real thing. It’s a very clever and sneaky way to obtain someone’s bank information, without too much risk on the part of the criminals.
However, vishing techniques come in many shapes and forms. The one described above is just the tip of the metaphorical iceberg. Here’s a good overview of the different types and techniques of vishing that are currently known to be used.
3. Deepnudes And Revenge Pornography
I’m sure you’ve seen them already, there’s no need to deny it. Because they’re already all over the web and also on those websites you enjoy watching frequently: naked pictures and videos of celebrities that never actually participated in such acts.
The name for this particular category of deepfake images and videos is ‘deepnudes’. It’s not difficult to do, anyone can do it with a simple app. While people enjoy creating images like that about celebrities, some of you would surely enjoy a nude photo of that cute blonde down the street. AI makes that fantasy a reality with DeepNude, an app that has since been pulled from the digital stores for breaking the regulations.
And it could happen to anyone. Revenge pornography is now easier than ever, with all the negative implications connected to it. Being targeted by deepfake pornography can potentially severely upend someone’s life. The effects are similar to that of the more traditional ‘revenge porn’ acts: personal shaming, harassment, blackmail, intimidation and silencing of victims being the worrisome result.
Manipulation is often the easiest when the victim is the most vulnerable. And it’s no longer needed to have your ‘amateur clip’ leaked on the internet. That photo on Instagram or Facebook is not more than enough to become a victim of such acts as well. When even a simple app can pull it off, imagine what someone with some more technological expertise could do with deepnudes.
People will go to great lengths to prevent such imagery from ever being publicized. There are plenty of known cases of revenge pornography that ended up in suicide by the victims of such publications. And even those that don’t take it that far will quite possibly suffer great psychological and social damage. It’s a theme that is certainly high up there in the ranks of deepfake manipulation and should not be underestimated.
4. Deepfake Identity Fraud
Grab your passport for a minute. That’s you on the image, and that’s your name, right? Of course it is, it’s your passport! Well, that’s not the case for all people on the planet. There are large groups of people that (for whatever reason) benefit from hiding their true identities.
The fake passport market is thriving on places like the Dark Web, where a single passport could be sold to anyone for surprisingly low amounts of money. And just like that, you can have your identity stolen by a complete stranger. Someone you’ve never even met before and that has no connection to you as a person whatsoever.
Identity fraud is a real problem, and it is ruining lives in ways you don’t even want to imagine. The issue will haunt an individual for many decades, without them even having to do anything. It’s been a problem for as long as the passport system exists, and it’s only getting harder to fight because of technologies like deepfake images.
Deepfake identity theft is a special type of identity fraud that could take one of three forms:
- The falsification of ID documents to access accounts
- Using the victim’s face and voice to change account numbers, passwords or authorized users
- Using the victim’s voice on a phone call to family and friends to extract funds from them
As you can see, identity fraud has some links to the aforementioned fraudulent vishing technique. However, the focus will be on passports and similar documentation, so the criminal can for example pretend to be someone else while traveling or doing a government check of some kind.
With deepfakes becoming more and more accessible to the wider public, trust in businesses and society will gradually start to erode. When you cannot be sure who is who anymore, how can you know that the person you let into your country, or the person you hire, is actually who they say they are? It’s a problem that will only get tougher to fight with the introduction of these easy to use deepfake technologies.
5. Fake News And Hoaxes
If you thought the Trumpian term ‘fake news’ was bad, deepfakes are the place where the truth goes to die. That’s how The Guardian famously introduced the wider public in their well-known publication on their website. And they’re not wrong, because how can we know that the things media are reporting actually happened the way they say?
We aren’t conspiracy theorists, but deepfake creators can essentially make reality into whatever they want. It’s a slippery slope where technology can make it look as if anyone has said or done anything. Without the wider public ever knowing that those things actually never happened.
Referring back to the introductory part of this article: all it takes is a single kid with bad intentions and a laptop with Wi-Fi connection. A hoax can be created by an edgy teen in a basement in rural Kentucky. Or by an underfunded and understaffed Russian troll factory in a suburb in Moscow.
The point is clear: fake news and media hoaxes might be a rare event today, but they will be a frequent occurrence as soon as deepfake software is more widely used by the broader public. A small bit of technological advancement, perhaps several years.
That’s all it’s likely going to take before the media landscape will be inevitably bombarded by such videos and images. It’s where media has to become very wary of any unverified news source, and has to set up countermeasures to verify sources even stricter and more thoroughly than they have always done in the past.
Deepfakes Are Here To Stay (Probably)
It’s probably redundant to say at this point, but government, industry and media outlets need to be prepared for the storm of ‘deepfakes’ that will inevitably come their way in the near-future. Deepfakes are already here, but still play a relatively small role in the world of manipulation.
That won’t last, looking at the pace at which the technology is improving and the exponential rate of adoption of deepfake apps and software around the globe. It’s a ‘digital virus’ that will likely make a global pandemic look cute. It’s important to start defending the public today and stay ahead of the curve, even though stopping the harm is – sadly – probably a lost cause already.
However, if we don’t put up a good fight, the impact that will be felt will be much greater. That’s why we need to stop deepfakes in their tracks, so manipulation by malicious actors can be reduced to a minimum. It won’t be easy, but we have to try. We created this problem, and we are the only ones that have the ability to stop it.
