The Deep World of Deepfakes
No, you don’t need to get your eyes checked. No, you don’t need to adjust your screen. You really are seeing the impossible. How could he say that? Why would she do that? Here’s a hint: they might not have. Deepfake technology can now create audio and video that is so convincing and realistic, that it can be almost indistinguishable from reality. Unfortunately that means that it can have very real and damaging impacts for companies.
Photos vs. Videos

In the past, photos were considered irrefutable proof. Now, most of us know how easy it is to manipulate a picture and are sceptical if something seems too fantastic. But videos? The majority of people still believe videos to be much more reliable evidence. After all, if they can see and hear someone doing something or saying something then “it must have happened.” However, it is imperative that we stop being so naïve.
Fact vs. Fiction
There’s a variety of ways that videos can be faked, including deepfake technology. Deepfake uses AI to piece together pieces of other videos and pictures, to create something brand new. This isn’t just taking an existing video clip out of context and making it appear to be something it’s not; this is the creation of something new altogether. Using audio clips of that person speaking, the technology can extract the precise sounds needed to form different words, and make it sound as though they’ve said something they never did. AI is literally stealing their voice, so it will genuinely sound like them and not just like a good mimic. Then, using hundreds or thousands of photos and video clips, Deepfake technology can learn the person’s mannerisms, and use them to construct a video of the person using their own body language and lip movements. The results can range from eerily similar, to indistinguishable, depending on how powerful the AI is, and how much content it has to work with. What does this mean for businesses? Unfortunately, nothing good.
Business Impacts
Deepfake technology is already being used to exploit companies. One of the first recorded cases of deepfake fraud involved AI generated audio of a company’s general director’s voice, and was used to wrongfully authorize the transfer of 220,000 euros. However, the possible implications go beyond just theft. The believable impersonation of business officials can lead to false claims that will damage a product or company’s reputation, or inseminate fake news about the officials themselves. In addition, it can be used to divert shipments of goods, order unwanted items, or cancel necessary ones, all of which can lead to financial losses.
The Threat is Enough
Perhaps the most frightening part about this technology, is that the events depicted in audio or video recordings don’t even have to be believable for very long. The authenticity of the video has no effect on the damage that it can cause because revealing a hoax is much more difficult when the casual observer cannot tell the difference between fact and fiction. To that end, the fake content doesn’t even need to be released, as the threat of release is often enough to extort and/or blackmail companies for malicious gain. Should you be worried? As with all potential threats, you should have a plan in place for how you will proceed if your company pulls the short straw and is the target of an attack.
This article was published in the
June 2020
edition of The TMC Advisor
- ISSN 2369-663X Volume:7 Issue:4
©2020 TMC Consulting