Artificial Intelligence

What is Deepfake and how do you identify fake content?

Publicado em 
19
/
03
/
2024
Home
/
Insights
/
Blog
/
What is Deepfake and how do you identify fake content?
Sign up for our Newsletter
Authors

With the continuous progress of technology, negative forms of use are inevitable. In the case of Artificial Intelligence, whose potential is immense, it is also exploited by malicious individuals to create what are known as”Deepfakes”, fake AI-generated content.

This misuse of AI affects individuals, companies, and even governments, causing financial, image, and reputational damage. It's no wonder that the Superior Electoral Court (TSE) approved, in February, 12 resolutions on the use of Artificial Intelligence in the 2024 elections, including banning the use of deepfakes.

The subject of artificial image editing came back to the fore after A photo of the Princess of Wales, Kate Middleton, be removed from news outlets on suspicion of “manipulation”. The princess later admitted that she edited the image herself.

In this article you will understand what deepfake is and how to identify this content so as not to fall into fake news. Check it out!

What is Deepfake?

Comparativo de Morgan Freeman gerado por IA e diretor que captou as imagensp
Director simulates Morgan Freeman's voice and face with deepfake

Simply put, deepfake content is fake videos, images, and audios that, using Artificial Intelligence and deep learning concepts, They look real. This technique is able to simulate expressions, voice, face shape, and create compelling fictional situations.

An example is the image above, which is an excerpt from a video by the Dutch filmmaker, Bob de Jong (at the bottom), in which he emulates the voice and face of actor Morgan Freeman with deepfake techniques. You can check out the full video here.

In this case, one person impersonates another, but there are several other ways of using deepfake in which only someone's voice or even an image is changed, as was the case of Pope Francis and the cold coat, which went viral last year.

Papa Francisco com uma jaqueta branca acolchoada
False image of Pope Francis that went viral in 2023

This problem is not so recent, but with the spread of Artificial Intelligence, it has become much easier to change any content. There is an enormous variety of applications and software capable of producing deepfakes with great speed and quality, especially when it involves audios, which is a full plate for scammers.

How do deepfakes affect companies?

As we mentioned, deepfake content is not restricted to celebrities or politicians, but also they can target companies. In this context, scams can have financial, reputational, and image objectives and also data breaches.

As an example, in Hong Kong, an employee of a financial sector company made a transfer of US$ 25 million after participating in a meeting forged with deepfake. The scammers recreated the voice and image of the company's chief financial officer and tricked the employee into completing the false business transaction.

That's why it's important to invest in cybersecurity to prevent fraud and digital attacks is essential. Here at Venturus, this is one of the biggest concerns when working on our clients' projects. Get in touch with our experts and find out how we can help you!

How to identify deepfakes

The good news is that there are already software and applications, some even free, that can identify content generated from Artificial Intelligence. And they work just like deepfakes, that is, they use the same technology, but aimed at the common good.

O Deepware, for example, is a software that identifies AI-generated videos. Just upload the desired video and the tool identifies if there has been any synthetic manipulation. But it should be noted that the result is not always accurate, especially when deepfake content is professionally produced.

Already Hive it is already focused on identifying textual or image content made by AI, even directly through the browser.

In addition to these and other software, it is also possible to identify deepfakes with a closer look and seeking other sources of information.

In the videos, for example, it is possible to notice some defects, even if they are discreet, such as:

  • lack of synchronization From the lips to the speech
  • Unusual movements of the blink of an eye.
  • Some details are missing Traits of the person

In the case of other types of deepfakes, such as audios, images, or texts, try to find another source of that news and avoid sharing content whose veracity you don't know.

Get ahead with the Venturus

Stay on top of the hottest discussions in the tech world with Venturus. Sign up for our monthly newsletter so as not to miss anything!

Artificial Intelligence
Artificial Intelligence

Contact one of our specialists will serve you:

Venturus - Campinas/SP - Brazil
Estrada Giuseppina Vianelli di Napolli, no. 1,185
GlobalTech Campinas Condominium
Polo II High Technology
ZIP CODE 13086-530
Venturus - Manaus/AM - Brazil
Rua Salvador, 440
Adrianopólis
ZIP CODE 69057-040
Venturus - Atlanta/GA - USA
1201 Peachtree Street NE,
Floors 1, 2 and 3,
Atlanta, Georgia
Zip code 30361
Please enter a corporate email
Best way to reach you*

Venturus needs you to provide us with your contact information to better talk about our products and services. By entering your details, you will be aware of the guidelines of our Privacy Policy.

Thank you! Your message has been received!
Oops! Something went wrong with sending your message.