Deepfakes: Protecting your business when you can’t trust what you see and hear
We already know that we can’t always trust what we read. Fraudulent texts, fake emails, and online misinformation are pervasive. And now, with the evolution of artificial intelligence (AI), it’s becoming increasingly difficult to trust what we see and hear in any medium.
Understanding deepfake technology
Deepfakes are videos, photos, or audio clips made with artificial intelligence to look real. The technology can be used to alter audio, video, and imagery from their original forms and make them appear convincingly authentic.
Deepfake technology can mimic or manipulate a person’s voice in audio recordings to say things that the person never truly said. In videos, a person’s face can be superimposed over someone else’s. And, a video of a person can be manipulated to appear as if they are speaking, moving, and gesturing in ways that they never actually did.
In this article, we provide more insights into deepfake technology, how it can be used to perpetuate deception and fraud, and actionable steps you can take to help prevent it from harming your business.
Deepfakes can be harmless or malicious
This technology is widely available and commonly used. Whether deepfakes are being made for harmful or harmless purposes, it’s important to recognize when the technology is being used, so you can continue to understand its capabilities as they evolve.
Harmless deepfakes
For fun, people can download apps that superimpose their faces on popular TV characters and put them into scenes from various episodes.
The movie industry employs the technology to edit character performances when the actors are no longer available, or to reverse-age characters for flashback scenes.
Retailers can use similar technology to allow customers to virtually “try on” sunglasses, clothing, jewelry, and more.
Malicious deepfakes
Disinformation agents can manipulate video footage of political candidates to show them saying degrading things or appearing intoxicated to sway voter preferences.
Bad actors can superimpose a person’s face over explicit video content then share the video to cause reputational damage or use it for blackmail purposes.
Criminals can fake a voice message using a CEO's voice and instruct an employee to release a large payment to a fake vendor account.
