Deepfakes are this century’s new approach to photo editing. Through the use of deep-learning artificial intelligence, computer imagery can be programmed to replace one person’s face with another. The results are realistic fake video and other digital content.
There is a variety of photo editing software available to accomplish these tasks, mainly open-source. The AI techniques are built in, and the user doesn’t need to figure out how the encoding and algorithms work.
Although deepfakes can be used in applications such as movies and gaming, the technology can also be used in a malicious manner. It has been used in politics and the entertainment industry to damage reputations or present misleading videos. Deepfakes can clone voices, faces, bodies, locations, almost anything.
There are a few ways to identify a deepfake. Sometimes the face animation is a bit unnatural. The focus might be blurred. Maybe the audio doesn’t match up precisely. As deepfake technology becomes more sophisticated it will become harder to spot.
Cybersecurity threats are evolving through the use of deepfake technology, giving attackers the ability to bypass biometric security, perform blackmail and social engineering scams. Known as synthetic identity fraud, a crook can combine real and fake information to create a new identity. So these miscreants can use AI to combine facial features from different people to fashion a new identity. The criminal can then use this new “person” to open fraudulent accounts or make purchases.
Industry experts are predicting a surge in the use of deepfakes for malevolent purposes. There is software available that combat deepfakes by using, ironically, artificial intelligence. AI and machine learning technologies need to be further developed in order to accomplish better verification processes.