Sirwin
Sirwin

DeepFakes: The Dark Side.


Technology is neither good nor bad, but everything depends on the use we give it. This is a phrase that is often repeated when talking about the scientific and technical advances that surround us, however, is it really true? Or is there something in human nature that inclines us to use technology for evil purposes?

In my previous article: DeepFake: Creating Them From Google Colab, explained in a simple way how to make DeepFakes, an example how artificial intelligence can be used to create fake videos or images that resemble reality. A comment from ZeroRequiem (whose articles I recommend: Publish0x, Bulb), made me reflect again on how this technology can have positive applications, such as entertainment, education or research, but also how it has a dark and negative side, such as deception, defamation, extortion or manipulation. What leads us to use DeepFakes to do evil? Is it a question of morality, psychology, sociology or physiology? What factors influence our decision to use technology for better or worse?

Unfortunately, as users we can also be unconscious victims of DeepFakes:

πŸ“Œ Identity theft: They have been used to impersonate someone and access their bank accounts, social networks, emails or online services. Earlier this year, a group of hackers used a DeepFake of the voice of a company's CEO to convince an employee to transfer $243,000 to a fraudulent account. There have also been reported cases of DeepFakes of faces used to apply for loans, credit cards or fake passports.

πŸ“Œ Falsification of evidence: Creating false evidence that incriminates or exonerates someone of a crime or misdemeanor. This year it made the news as a man was accused of murdering his wife, but he presented a DeepFake video as an alibi, showing that he was somewhere else at the time of the crime. There have also been reported cases of DeepFakes used to fabricate false testimonies, false confessions or fake crime scenes.

πŸ“Œ Disinformation and propaganda: Spread false news, rumors or propaganda that affect public opinion, politics or national security. A few months ago a DeepFake video went viral showing the president of the United States saying that he was going to launch a nuclear attack against North Korea. There have also been reported cases of DeepFakes used to alter the speech of political, religious or social leaders.

πŸ“Œ Violation of privacy and consent: They are used to create intimate images or videos of people without their consent, either to humiliate them, blackmail them or sexually exploit them. A website was discovered where a DeepFake porn could be requested from anyone in exchange for money. Cases of DeepFakes used to create revenge porn, sexual harassment or sexual extortion have also been reported.

To give you an idea of how far we are exposed, I tell you about a news story that went viral just a few days ago, and that was related to the YouTuber with the most subscribers in the world, Jimmy Donaldson, better known as MrBeast. On October 4, 2023, he reported a DeepFake that was circulating on social networks in which he appeared to be offering an iPhone 15 for just two dollars. Although I cannot corroborate an exact figure, it is estimated that there were thousands of users deceived by this scam.

https://twitter.com/MrBeast/status/1709046466629554577?s=20

https://youtu.be/VI_S-FBfz1g?si=PN6MFcclQSe6QUKu

https://youtu.be/QSesQdF2y40?si=tKvFz5rGVMxeBrb2

To avoid being a victim of a DeepFake, it is important to be skeptical of the videos you see on social networks. If a video seems too good to be true, it is better to be suspicious. It is important to check the source of the video, if it comes from an unknown source, it is better to avoid it. Also, be careful with links shared in videos, if the link seems suspicious, it is better not to click on it.

To detect a DeepFake, with current technology, you have to be very good observers, although we can look at some details that still reveal its falsehood:

πŸ“Œ The number of blinks, for example, people blink about 15 times per minute, but DeepFakes usually have fewer or very exaggerated blinks.

πŸ“Œ The face and body, currently DeepFakes usually focus on the person's face, but neglect the rest of the body, which may have a different shape or size than the original.

πŸ“Œ The length of the video is usually short, just a few seconds, because they require a lot of time and resources to create. For example, the video I made for the previous article has a size of 94.7 Mb, lasts just 16 seconds and took approximately 2 hours to create.

πŸ“Œ In addition, DeepFakes may have synchronization problems between audio and video or a voice that does not match the age, gender or accent of the person.

πŸ“Œ Also shadows and skin tones tend to have inconsistencies in lighting, skin colors and can vary from frame to frame.

πŸ“Œ In soft or blurry areas, DeepFakes have difficulty reproducing the natural movements of the mouth, eyes, hair or clothing and look distorted or pixelated.

Another way to detect DeepFakes is to use artificial intelligence tools that analyze images, videos and audio tracks to look for signs of manipulation. Some of these tools that I found on the internet are:

πŸ“Œ TensorFlow and PyTorch: Two examples of free tools that use deep neural networks to detect DeepFakes. They can be used to analyze images, videos and audio tracks to detect signs of tampering. Users only have to upload examples of real and fake content to train a detection model that is capable of differentiating between the two.

πŸ“Œ Deepware: An open source technology primarily dedicated to detecting AI-generated videos. It has a scanner to which the material can be uploaded to find out if it is synthetically manipulated. Like other DeepFakes detectors, deepware models look for signs of tampering on the human face.

πŸ“Œ Sensity: A tool specialized in detecting DeepFakes and images generated by artificial intelligence (AI). It uses a combination of computer vision, machine learning, and forensics to identify and track the spread of DeepFakes on the internet.

πŸ“Œ Hive: A platform that offers verification and moderation services for multimedia content. It uses a network of more than a million people who manually review suspicious content, as well as AI algorithms that detect anomalies and fakes.

πŸ“Œ Illuminarty: A mobile application that allows users to verify the authenticity of images and videos they receive or view on social networks. It uses facial analysis techniques, optical character recognition and metadata to determine whether the content has been altered or not.

You must keep in mind that this DeepFakes technology is constantly evolving and can improve its quality and realism, therefore, I recommend that you always verify the source and credibility of the content you see on the internet, and that you do not be fooled by the appearances.

As you see, DeepFakes are an invisible threat that can ruin our lives, their impact can be devastating for many due to the growing threat to privacy and security. It is important to be alert and verify the authenticity of the images and videos you see on the internet. It is also necessary that whenever you protect your identity and privacy online, you do not share sensitive personal information or images that could be used against you, such as high-resolution photographs or personal details. Remember that DeepFakes are not just innocent fun, but a dangerous tool that can cause a lot of damage.


TOOLS, PLATFORMS & APPLICATIONS

πŸ’²Β Bulb,Β Publish0xΒ - Earn daily Crypto, NTFs or Money for reading or writing articles and interacting with publications.

πŸ’² AddmeFast - Earn daily Crypto. Promote and increase the sources of traffic, visibility, reach and reputation of your social networks.

πŸ’² Bitrefill - Living with crypto, a philosophy of financial freedom. Travel, play, eat and live with BTC.

πŸ’² QuantFuryΒ (Invite Code: JRRU2593) - Trading - Join using my invite code: JRRU2593 and we will both receive a free share like AAPL or UBER, or crypto like BTC or ETH (up to $250). Trade and invest with no commissions or borrowing fees at real-time spot prices from the NYSE, Nasdaq, CME, Bats, Binance and Coinbase exchanges. With a good marketing management you have the possibility of obtaining passive profits without operating in the market.

πŸ’² StormGainΒ - Trading - They can startΒ without investment, capital is acquired for free with the Bitcoin Cloud Miner.Β 

πŸ’² BingX - Trading - Called "The People's Exchange", it places a strong emphasis on social trading and offers its clients extensive features: new user rewards, demo account, high leverage, spot trading, standard and perpetual futures, grid trading, copy feed , etc.

πŸ’²Β CoinEX, KuCoin - Trading - They offer different bonuses.

πŸ’²Β Bitcoin Spark - ICO - Initial Coin Offerings.


Author's Note: The opinion expressed here is not investment advice, is provided for informational purposes only, and reflects the opinion of the author only. I do not promote, endorse or recommend any particular investment. Investments may not be right for everyone. Every investment in the market and every trade you make involves risk, so you should always do your own research before making any decision. I do not recommend investing money that you cannot afford to chair, as you could lose the entire amount invested.

πŸ’–πŸ’–πŸ’– Contribute To The Author πŸ’–πŸ’–πŸ’–

How do you rate this article?

44


CryptoEntrepreneurs
CryptoEntrepreneurs Verified Member

Trader, entrepreneur, freelancer, enthusiast and promoter of globalization, adoption and understanding of cryptocurrencies and new information and communication technologies (ICTs)


A bit of everything ...
A bit of everything ...

β€œEveryone has their own forms of expression. I think we all have a lot to say, but finding ways to say it is more than half the battle." Criss Jami (existentialist philosopher, poet, essayist, musician, singer, designer and lyricist). | "Everything we hear is an opinion, not a fact. Everything we see is a perspective, not the truth". Marcus Aurelius (Roman Emperor).

Send a $0.01 microtip in crypto to the author, and earn yourself as you read!

20% to author / 80% to me.
We pay the tips from our rewards pool.