AI deepfake apps are becoming increasingly prevalent, and the implications of their use are both far-reaching and alarming. From the ability to track down images of people or replicate their voices, these AI deepfake apps are leading us toward a future where video evidence may no longer be reliable. Let’s explore the threats and realities that AI deepfake apps present, as well as examine how we might be able to protect ourselves from them.
What are deepfake apps?
Deepfake apps are computer programs or websites that use artificial intelligence (AI) to generate realistic-looking videos, images, and audio of people. These deepfakes are created by feeding the AI system with hundreds of images, videos, and audio recordings of a person, which the system then uses to “learn” their facial expressions and vocal patterns. AI then uses that data to create a realistic representation of someone that looks and/or sounds like the real thing.
Deepfakes have become increasingly common in recent years due to the advances in AI technology. For the most part, they’ve been used to create content. You can find interviews of people like Bill Hader doing impressions edited with a deepfake face of the person he’s impersonating.
It’s great fun and showcases the light-hearted entertainment purpose of AI. In Disney+’s hit show “The Mandalorian,” deepfake technology superimposed actor Mark Hamill’s de-aged face on a stand-in actor when the character of Luke Skywalker appeared in an episode.
However, while some deepfakes are harmless, there is potential for them to be used in nefarious activities. While deepfake technology isn’t completely flawless yet, its potential harm is evident.
The underlying threat
The most obvious threat is to privacy and security. AI deepfakes could be used to create false identities or false images of people, which can then be used for malicious purposes. AI deepfakes could also be used to manipulate information, spread false information, and even threaten public safety.
Another potential threat is to the integrity of evidence. AI deepfakes can be used to manipulate video and audio evidence, making it difficult to tell whether a piece of evidence is real or not.
Video evidence might not be enough in a world where people can manipulate the faces of those in a video; adding voice emulation just makes things even more complicated. This could have serious implications in criminal cases and other legal proceedings, as well as for investigative journalism. Not to mention, deepfakes could wreak havoc on people’s lives in the social media court of public opinion.
There are several AI voice replicators that only need a short recording of people’s voices to emulate them with surprising accuracy. Considering a lot of recorded audio is already somewhat distorted, it wouldn’t be too hard for people to believe a deepfake audio, even if it doesn’t sound exactly like the person it is emulating.
AI deepfakes could be used to target individuals and cause damage to their reputation. AI deepfakes could be used to create malicious images or videos that appear to show an individual in a compromising or embarrassing situation.
There is already a growing issue with deepfakes and the internet’s infamous obsession with pornography. Deepfakes could have severe psychological and social effects on people, and we aren’t really prepared to prevent their impact.
What can we do?
The availability of image-tracking software and AI that can pull your data from the internet is extremely concerning. With the number of pictures people post of themselves on social media, malicious actors could have access to thousands of pictures of almost anyone online.
This issue highlights the importance of online privacy. Do your best to protect your data in general, and make sure there aren’t too many images of you floating around online. While that sounds counter-intuitive in the age of social media, it could pay off down the line as AI continues evolving into the world’s dominant technology.
