While technology advances at breakneck speed, misinformation and disinformation are keeping pace. Deepfakes are getting quite common on social media platforms like YouTube, Facebook, TikTok, and Instagram. Recently, a deepfake video of the Ukrainian President Volodymyr Zelensky was circulating on social media sites where he seemed to tell Ukrainian soldiers to surrender to Russian soldiers. Even though the video clearly appeared to be a deepfake, it still posed the question about how this technology can be used to spread false information, especially if the media gets blindsided as well. A similar situation unfolded recently amid heightened tensions between India and Pakistan, where manipulated videos and misleading posts circulated widely, falsely portraying cross-border military actions. These viral pieces of content not only inflamed public sentiment but also risked escalating conflict based on fabricated narratives.
Deepfakes didn't even start with politics, rather, they started with deepfakes of celebrities like Taylor Swift and Gal Gadot. Soon, many companies started using them, and one company even allowed people to animate the pictures of their deceased loved ones to bring them to life. Nowadays, deepfakes have become so advanced that it is hard to differentiate between what's real and what's fake. In this article, we will explore deepfakes a bit deeper and how people and organisations can identify them to keep themselves safe from being fooled.
What is a Deepfake?
A deepfake is a fake video, image or audio created by AI that seems real, and the technology used to create it is called deep learning. The
technology and
tools make people do or say things they never actually did, and many public figures become victims of it.
Deepfakes have been growing a lot since 2018, and now they have reached a peak in numbers, with
85,000 harmful deepfakes by the end of 2020. Deepfakes are also harmful because they can spread fake news, support unethical political goals, and be used for revenge as well. If you want to create powerful and realistic deepfakes, it takes skilled experts and powerful computers, but now people can even do it without having any knowledge because of cheap or free apps. This is because of advancements in AI and cloud technology, and it's getting harder to separate fake from real.
Are Deepfakes Illegal?
Now that a lot of people are using deepfakes, another question arises about the legality or illegality of deepfakes. The answer is that it isn't completely illegal to create deepfakes, but it depends on how they are being used. Many people use deepfakes for entertainment, and they are quite harmless, but the ones used to harm and exploit someone or spread misinformation can become illegal.
In the EU, the rules around
deepfakes are strict, and regulations like AI laws, GDPR, disinformation policies, and copyright rules tackle deepfake issues legally. However, none of these regulations directly address deepfakes, and it is still unclear if deepfakes can be used in courts as evidence.
Israel recently introduced a new law that requires all edited images to be labeled as such, and this can soon be applied to deepfakes as well. Other countries are also trying to deal with deepfake issues on their own, but none have come up with any clear regulations and laws about deepfakes.
Deepfakes: How Do They Even Work?
We all now know that deepfakes work by using artificial intelligence, but it's actually the neural networks that make deepfakes mimic how someone looks or sounds. AI models are trained on data sets of thousands to millions of images, audio clips, and videos, and they can learn to mimic the voice, facial expressions, and movements of a person with accuracy.
When it comes to face swapping, the AI creates a digital version of someone's face and then overlaps it with someone else’s, which perfectly mimics the individual’s facial expressions.
Voice cloning also works in a similar way with AI analyzing the audio recordings of a person's voice and then generating a new audio that sounds exactly like what the person could have said.
The more visual and audio data the AI has, the more realistic the results become, and as technology is improving, it is becoming harder for people to differentiate between real and fake. There are different apps for
deepfakes like FaceApp, which is mainly used for fun but can still make real and convincing videos of people, even if the user doesn't have any technical skills.
The Purpose of Deepfakes
Deepfakes aren't always used for malicious purposes, as they can also be used for educational, creative and empowering causes as well. They can be used in a positive way as well as in a negative way.
Positive Uses of Deepfakes
In positive settings, deepfakes are being used in films to de-age actors or bring back performances of people who have passed away. They are also being used in voiceovers, parodies, and e-books for entertainment purposes and to make the storytelling more engaging and interesting.
Some teachers are also using deepfakes in educational settings for historical figures and to bring them ‘to life’ to make it more fun and engaging for students. Deepfakes are also being used for marketing, virtual exhibitions, and presentations. Even criminal investigators are using deepfakes to enhance their communication and analysis.
Negative Uses of Deepfakes:
On the other side, deepfakes are also being used in a negative and exploitative way. Deepfakes can be used for identity theft, fraud, and to blackmail people about things they never said or did. Deepfakes are also spreading fake news that looks believable and real, especially when there's a conflict or a political event.
Deepfakes are also being used for warfare manipulation, especially seen during the Ukraine war when the fake video of the President was circulating. There was another example of deepfake being destructive when actor Jordan Peele made it appear that Barack Obama was calling Donald Trump a name by using a deepfake.
Spotting a Deepfake
Even though deepfakes have become more real than ever, there are still some signs that can help you spot them even if you aren't a tech expert. The following are some ways to detect a deepfake:
1- Check the Source:
Before believing anything you see on the internet, always ask the question: Where does the video or the image come from?
If it is from an unknown or suspicious account, be cautious and do not believe its authenticity it. Check if it's a fan or parody account or a trusted news outlet, and always look for reliable sources.
2- Take a screenshot
Take a screenshot of the image or the video and run it on Bing Images or Google Images. This can help you trace whether the image is real or what the original version of it is. Also, check the sources from which the images have been posted.
3- Fast Check Yourself
Quickly check if the deepfake is being reported by credible news sources. If no trustworthy news outlet covers the story or event, it is probably fake. Trusting yourself and your gut feeling when it comes to deepfakes is also important.
There are also some other ways that can help you detect a deepfake:
● Look for visual cues like unnatural head or body moments and weird lightning or color.
● Notice any strange eye movements, unnatural blurs, like blinking too hard or their eyes moving in a strange manner.
● Look for facial expressions and notice if their face is aligning with their emotions well or if there are no expressions at all.
● AI always struggles with fine details like teeth and hair, so take a keen look at them, and if they are too perfect, it can be a deepfake.
● Listen to the audio carefully and hear if it feels off or mismatched, and if there are unnatural silences. Look at whether the mouth is matching the words perfectly or not.
● Last but not least, always trust your instincts and confirm the reliability of the video before sharing it.
Practical Tools and Resources to Detect Deepfakes
There are several tools available today that can help you identify deepfakes and verify the authenticity of videos and images. For example,
Deepware Scanner allows users to upload videos and check if they are manipulated using AI detection models. Another great option is
InVID, a powerful browser plugin that breaks down videos into key frames and helps analyze images and videos shared on social media.
Sensity AI offers advanced deepfake detection services used by companies and researchers to spot manipulated media.
Reality Defender uses real-time AI to detect and flag deepfakes on websites and social platforms. Combining these tools with trusted fact-checking sites and
critical thinking will help you avoid falling victim to fake content.
Effect of Deepfakes on Companies:
Deepfakes are a serious threat to companies because they are hard to spot and can be used in fraud. Cybercriminals often target celebrities, politicians, and businesses by making their harmful deepfakes. For example, they can make a deepfake of a company’s CEO saying something controversial or leaking sensitive information, and this can lead to a loss of finances and damage to the reputation of the company.
In 2020, a criminal used a deepfake voice of the CEO of a company and tricked a bank director in Dubai to
approve the transfer of $35 million. The voice sounded so real that the director believed it was real, and many similar scams have come to the surface now, where criminals target people for money using deepfakes.
Companies must stay alert as deepfakes become a serious security risk. Training employees to recognize suspicious videos or audio is essential, especially for executives who might be impersonated. Implementing strict verification procedures for sensitive communications—like confirming unusual payment requests via multiple channels—can prevent costly scams. Many organizations are also adopting AI-powered security tools to detect manipulated media before it reaches internal systems. Staying informed and having clear policies around digital content helps reduce the chances of falling victim to deepfake fraud.
Keep Yourself Safe From Deepfakes
Deepfakes began as something simple, but now they have become a global threat and can be used to make convincing fake videos and images of people and events that can become problematic. Now, anyone can make a deepfake, which means no expert is needed for it, so now it has made everyone unsafe in this digital world.
If you come across a video or image that seems suspicious, avoid sharing it until you verify its authenticity. Report the content to the platform where it appeared, such as YouTube, Facebook, or TikTok, which have policies against manipulated media. Use the verification tools mentioned earlier to analyze the content or check if credible news sources have covered the event. Sharing unverified deepfakes only helps spread misinformation and can cause real harm, so a cautious approach protects both you and others.
Deepfakes are becoming realistic, and spotting them is becoming a challenge, but human awareness and using the right technology can make it stop before it becomes a big threat.
Image: DIW-AIgen
Read next:
ChatGPT Usage Statistics: Numbers Behind Its Worldwide Growth and Reach
by Arooj Ahmed via Digital Information World
