5 Ways To Detect And Recognize Deepfakes (As An Average Person)
With deepfake technology quickly becoming a part of our daily lives, it will be more important than ever to be able to detect deepfakes yourself. But how does one do this, exactly? Isn’t the whole premise of a deepfake to go undetected by the broader public?
It’s a complex problem, but the methods one can employ to defend against false media content created by deepfake tech are relatively easy to implement. It might seem counter-intuitive to say this, mainly because the average media consumer can usually be easily fooled by a realistic deepfake video.
This article will explore the things you, and everyone around you, can do to navigate your way towards the truth in a world filled with fake media content. It likely won’t be a 100% success story, and it’s not a miracle cure either, but the strategies described in this article are the best ammo we’ve got to combat a world filled with content that can be described as deepfake videos, images and audio.
How Can An Average Person Detect Deepfakes?
The idea behind deepfakes is that they’re practically undetectable. Therefore, it will be impossible for an average person to systematically detect deepfake content without appropriate deepfake detection software.
However, this does not automatically mean we have to accept our new reality. There are ways for the average human being to know if something they’re viewing or listening to is fake. It’s all got to do with logical thinking and deductive reasoning.
We have to shift our collective mindsets a little bit and start teaching critical thinking skills in the mandatory education system. The children of today will be confronted with deepfake media content much more than the current generation of adults, assuming that the progression of Artificial Intelligence (AI) will influence our daily lives more and more over time.
This is exactly why it is now more important than ever to start educating the broader public about the need to detect and fight the presence of malicious deepfakes. We are talking about the type of deepfakes that are intended to mislead the public to further an agenda. We’d like to share five methods and strategies to start doing this right now.
Methods To Protect Yourself From Deepfakes
Looking at the practical strategies an average human being could employ to start detecting deepfakes, the toolbox is limited. We’d all have to start being critical thinkers and not take everything we are being told for granted. That’s a difficult task to accomplish.
Sometimes, clues will be obvious. After all, not every deepfake media video is 100% foolproof. This means that some deepfakes videos will give away the fact that they’re not real. They tend to give away small clues, such as facial glitches or odd looking lighting. These are common giveaways that most deepfake detection software tools exploit as an effective detection strategy.
For the average Joe, this won’t always be relevant. The first lines of defense will have to come from a critical eye that knows how to consume and digest new facts picked up in the media, whether that is traditional media (such as a news TV channel), or more modern media channels (such as social media on the internet). This thought process shapes the strategies the average person can adopt in the fight against false media content.
1. Question Everything
Why question everything? Good job, you asked a question right away. That’s exactly what we want! Doing something as simple as that, you’re already on your way to think critically. Questioning anything and everything will keep you on your toes at all times.
Questions that can always be used will generally come down to: Why do people say the things they say? Does it make sense for this person to make a claim like that? Would this person actually do or say something that is outside of what they normally talk about?
Whenever you see something on television, in the newspaper, or on social media. First ask yourself this question: ‘Is it real, or not?’
A healthy dose of skepticism can weed out a lot of the more obvious fake media content out there. Chances are that most fakes are actually quite easy to spot, once you know what to look for. A simple training for school children or students of all ages, could make a world of difference in shifting the behavior towards consuming content.
However, one must be very critical about this. It’s a slippery slope to start being critical about everything you hear, simply because you do not agree with the statements that are being made. This is especially risky once the statements start to be more and more political in nature. That’s where a divide is easily created. We must also be wary for more extreme cases, in which someone believes in their own truth so much, it starts leaning towards conspiracy theory thinking.
Unfortunately, not every person will be able to master critical thought in a productive way.
Solving this issue will be one of the main problems in a future society filled with deepfakes. If nothing is real, isn’t it much easier to create your own reality based on ‘facts’ you see as the absolute truth? Who says your absolute truth is the real truth? Isn’t it heavily influenced by your personal experiences and statements you picked up from people around you?
As you can see, the solutions aren’t as straight-forward as one might assume. Nevertheless, it is good to at least consider the possibility that the media content you consume might be a deepfake. Finding the balance within that, is a topic of debate for another time.
2. Verify With Multiple Sources
Next to being critical, verifying a story is as simple as a Google Search. Don’t consume news content from a single news source, but consume media content from news sources with different political orientations. Balancing out both news sources will most likely lead you towards a (relative) truth.
Even better would be taking the classic journalism approach. Going straight to the source of the news to verify a story is now easier than ever with the rise of social media platforms.
Check out what the person that is the center of the news scoop has to say about it. Do they refute claims made in the media? Do they confirm what is being said? You obviously cannot pick up the phone and call them, or pay them a visit in their homes. But if the source of the news makes a public statement about the news that is currently circulating, chances are that truth can be found in these types of statements.
Same story for a conflict between two parties. Check out the official statements made by the parties themselves (e.g. two companies fighting over a patent might make a public statement about it). The two claims will oppose each other, and using them as a source allows you to form an informed opinion about the situation.
So in essence, the verification of sources is about seeing the full picture. You don’t limit yourself to a single piece of a puzzle, but carefully put the pieces together and only after an image emerges, you have the ability to make an informed decision. No deepfake, regardless of how realistic it will be, will be able to fool an informed individual.
3. Look For The Minor Flaws
Deepfake detection software exploits this method to the fullest. Deepfakes aren’t always as perfect as people make them out to be, at least, for the time being. Looking for minor flaws in deepfake videos is often also possible with the naked eye. A deepfake will give away its true self one you start noticing the following aspects in the videos themselves:
- Video blurriness: The overall video quality is the first major giveaway. The production of a deepfake video takes a lot of computing power. Users of such software will generally reduce rendering time of deepfakes and keep them as blurry as possible, as to not give away the fact that the changes are fake. Most notably, the face will be blurry compared to the rest of the video environment. But usually the video quality overall will already be relatively bad to begin with. It’s easier to hide a fake if you can’t really make out any details to begin with.
- Face distortions: In most deepfake videos, you will be able to see some type of distortion around the edges of the face. This can be an issue with the light, or for example a mouth that will not feel natural in its movements. That immediately brings us to the next point of bad audio synchronization.
- Bad audio sync: The mouth will often move independently of the audio in deepfake videos. This could be shrugged off as a technical issue, but can also be a giveaway clue of the fact that the video has been tampered with at some point in the past.
- Odd eye blinking: Next to the mouth, the eyes are the easiest place to notice something might be off about a video. Eye blinking is a dead giveaway in this regard. Humans have a tendency to blink their eyes at a certain interval, as well as during certain moments of talking. If anything about the eye blinking feels unnatural to you intuitively, chances are that we are dealing with a deepfake video.
- Bad emotion intonations in the voice: Another clue can be found in the way someone talks. The emotion in the voice is sometimes difficult to realistically implement in a deepfake, making it relatively easy to spot. Also, it can be that the body movements of the person will not match up with the way emotion is used in the voice. That’s another way in which intonation of the voice can tell you that you might be dealing with a deepfake video.
This is just a small handful of ways one could detect a deepfake video without ever having to use a deepfake detection software tool for this. The human brain has evolved to detect very minor distortions in faces, as we use facial recognition in our everyday lives (and from an evolutionary perspective, we used to tend to use those skills for survival reasons).
4. Focus On The Context
The next simple thing you could do is to think about the topic of the content you’re consuming. What video are you watching? Is the topic controversial? Could there be a reason anyone wants to tell you something that involves a ‘hidden agenda’?
We aren’t talking about conspiracy stuff here, just controversial topics. Anything political is a given. But also powerful people in the world of entertainment or business could be an easy target for a deepfake video.
The context matters, because it tells you if it even makes sense to create a deepfake video about. If someone has a probable reason to mislead the public and sway public opinion, there will likely be a chance of a deepfake lurking in the shadows somewhere. This is especially true for controversial political content during election times. Not just on TV, but mainly on social media channels like Twitter or Facebook.
5. Download Or Use A Software Tool
The last resort is to download a software detection tool yourself. These tools are often open-source software freely available on the internet, so anyone can look them up and use them at will.
For example, the reverse image search functionality on Google is a great way for people to see if they’re being misled by a deepfake image. Is the image used anywhere else? Can you see what types of websites are using the image and in what context? Can you draw any conclusions from that, is the image suspicious or not?
For videos, this will already be a bit more difficult to do. There are a few corporate solutions that allow you to place fingerprints inside videos to verify authenticity, such as Amber.
For YouTube content, you could use a website like Deepware to see if you’re dealing with a potential deepfake video. The tool is currently still in open beta, but the full version is expected to launch soon. At the time of writing, it is the most comprehensive quick detection verification tool for the broader public that can be used on the internet.
Detecting Deepfakes: An Important Skill To Have
We have provided you with five quick strategies you can use today to start detecting deepfake videos, images and audio quickly as a beginner. Most of the methods explained in this article involve some type of common sense, or a method of deduction where a few audiovisual hints can indicate if a deepfake risk is realistic or not.
As time progresses, the realistic nature of deepfakes will only get better. With the widespread availability of deepfake software to create your own deepfake content, these types of videos will start popping up more and more in the media landscape.
Therefore, we strongly believe that educating people about the dangers of deepfake content is critical. At a young age, children should be taught about the existence of such videos and how to think critically about the risks involved.
Useful software tools such as Deepware show us that a flood of deepfake detection tools on the internet is likely to be created and released in the near future. When this will happen will largely depend on the speed with which deepfake content is going to spread on the internet in the next few years. More deepfakes will obviously increase the need for such tools to be more commonplace.
With more deepfake content being uploaded to video content websites like YouTube every day, tools like these will be the difference between finding out what the truth might be, and being fooled by a seemingly simple bit of content on the internet. We recommend actively implementing the strategies explained on this page to protect yourself from the inevitable influence of deepfake content, especially on the internet.
Knowing what types of clues to look for in a video that might indicate that a deepfake video could be present will be key. Especially when you’re a big consumer of political content on either TV or social media platforms, it will be crucial for you to start implementing the advice right now. Deepfake content is already a widespread phenomenon in the political sphere, especially with elections around the corner. There are plenty of countries with ulterior motives and endless resources that are willing to stir things up a bit before the next American President will be elected. Defending a fair and honest election will be in the hands of the public. Start educating people today about deepfakes, in order to safeguard a better tomorrow for this generation, as well as the next one.