The Dangers of AI-Generated Fake Images and Voices

I’ve written kind of a kilometer (yes I use metric) about AI, and I’m not nearly done. This time I want to talk about when AI gets used to create fake images or even fake voices. I’m going to start with an example. These are two different pictures of Henry Cavill. Can you spot the AI generation?

AI is rapidly advancing, and with it comes the ability to create increasingly realistic fake images and voices. This technology, known as deepfakes, has the potential to be used for a variety of purposes, both good and bad.

What are deepfakes?

Deepfakes are created using a type of machine learning called deep learning. Deep learning algorithms are trained on large amounts of data, such as images or audio recordings. Once trained, these algorithms can be used to generate new images or audio that are very similar to the original data.

In the case of deepfakes, the algorithms are trained on images or audio recordings of a specific person. The algorithm then learns to identify the unique characteristics of that person’s face or voice. Once the algorithm has learned these characteristics, it can be used to generate new images or audio that look or sound like that person, even if they never said or did the things that are being depicted.

What are the dangers of deepfakes?

Deepfakes can be used to spread misinformation, damage reputations, and even commit fraud. For example, a deepfake could be used to create a video of a politician saying something they never said, or to create a fake audio recording of a businessperson making a fraudulent statement.

Deepfakes can also be used to create non-consensual pornography, also known as revenge porn. This is a type of pornography that is created and distributed without the consent of the person in the images or videos. Revenge porn can be extremely damaging to the victim’s reputation and emotional well-being.

How can we protect ourselves from deepfakes?

There is no easy answer to this question. Deepfakes are becoming increasingly sophisticated, and it is becoming more difficult to tell them apart from real images and audio. However, there are a few things that we can do to protect ourselves:

Be critical of what you see and hear online. Don’t believe everything you see or hear, especially if it seems too good to be true.
Be aware of the signs of deepfakes. There are a few telltale signs that a video or audio recording may be a deepfake. For example, the person’s lips may not be moving in sync with the words they are saying, or there may be glitches in the video or audio.
Use fact-checking websites. Some websites can help you to verify the authenticity of online content.
Report deepfakes to the authorities. If you see a deepfake, you should report it to the appropriate authorities. This will help them to track down and prosecute the people who are creating deepfakes. Please, don’t do this with me, I am just using some imagery for the purpose of education.

What is being done to stop deepfakes?

There are a number of companies and organizations that are working to develop technology to detect and prevent deepfakes. For example, some companies are developing algorithms that can identify the telltale signs of deepfakes. Other companies are developing software that can help to create more realistic deepfakes, in the hope that this will make it more difficult for bad actors to use deepfakes to spread misinformation.

The United States government is also taking steps to address the threat of deepfakes. In 2019, the Department of Defense released a report on the potential dangers of deepfakes. The report called for the development of new technologies to detect and prevent deepfakes, as well as for the creation of new laws to punish those who create and distribute deepfakes.

The future of deepfakes

Deepfakes are a powerful technology that has the potential to be used for both good and bad purposes. It is important to be aware of the dangers of deepfakes and to take steps to protect ourselves from them. However, it is also important to remember that deepfakes are still a relatively new technology. As technology develops, it is possible that we will find new ways to use deepfakes for good.

In the meantime, it is important to be vigilant and to use critical thinking skills when consuming online content. If we can do this, we can help mitigate the dangers of deepfakes and ensure that this technology is used for good.

Oh, and the example at the beginning of this post; they are both AI-generated.