Deepfakes will be the most disruptive technology of the next decade.
In more ways than you think.
Deepfakes are media created by artificial intelligence algorithms. They typically refer to the impersonation of another person through video and/or voice. The technology has been around for a number of years and some notable, viral examples include:
Celebrity impressions by Jim Meskimen
Jordan Peele's impersonation of President Barack Obama
Popular YT channel 'Ctrl Shift Face' remakes clips of popular movies with different actors
Recently there was a highly publicized scandal where a Twitch streamer was caught browsing a site that offered paid porn deepfakes of popular female streamers. The controversy has sparked a heated debate about the ethics and implications of this technology, as well as a lawsuit against the site's creator. However, it is surprising that the potential for disruption seems to have been underemphasized by the media, especially given their tendency to eagerly fear-monger about threats to civilization. While the technology is imperfect and still marching through the uncanny valley, there is no reason to assume that it won't soon become indistinguishable from the real thing and easily accessible.
I believe this is the calm before the storm and the consequences for the next 5-10 years are understated.
Here are some areas that will be affected by deepfakes:
1. Existential anxiety and abuse
When photography was first invented, some people (especially in indigenous cultures) feared being photographed, believing it to be the imprisonment of their soul. Doppelganger tales can be traced back to folklore and mythology. There seems to be something innately threatening about the idea of our identity being copied or stolen.
Ordinary people will soon be confronted with a reality in which it will be trivially easy to create deepfake pornography of them or to place them in other degrading scenarios while perfectly mimicking their faces, voice, and body shape. Unfortunately, this means that flawless deepfakes will inevitably be used as a tool of abuse. It is difficult to overstate the potential impact on teenagers, who may experience even more traumatic forms of bullying in school as a result.
2. Social media privacy movement
Deepfakes require source material - videos, images, audio of the person they are supposed to be imitating. Therefore once it becomes clear through high-profile cases that the technology exists and that highly realistic deepfake porn is fairly easy to create, people will be more hesitant to post their faces online. Parents will certainly be much more concerned about sharing photos of their children. A public discussion about increasing the minimum age requirements for social media will likely follow soon after. Setting profiles to completely private will become much more common, perhaps a default. A popular movement is likely to emerge encouraging people not to upload pictures or clips online (which spells trouble for sites like Instagram or Tiktok).
3. Disinformation and polarization
Deepfakes arrive at an unfortunate time as trust in media globally is at an all-time low. For this tendency not to spiral even more out of control, journalists must be diligent about verifying the authenticity of the audiovisual materials they share.
People will obviously become much more susceptible to manipulation. Few people read beyond the headline; it's hard to expect them to investigate every viral clip that makes its rounds on social media. Clever fakes will take original material and make very subtle changes (e.g., inserting a crucial word into a statement that dramatically changes its meaning).
As a society, we are already deeply polarized on issues such as politics. This can be exacerbated when real evidence to the contrary can start being dismissed by some as ‘probably deep fake’.
Social media platforms have a huge responsibility to create tools that automatically verify the authenticity of content as it is uploaded. The question is whether these solutions will be able to keep up with a technology that is expected to continuously improve. And more importantly, how do we approach the audio/video form when we get to a point where it's impossible to determine its authenticity?
4. A less effective legal system
Many cases where the crucial piece of evidence is audio and/or video will often be contested and need to be verified by authentication experts. We are very close to this point as showcased by the recent Andrew Tate case where some of his fans tried to argue online the incriminating voice messages were deepfakes.
If we reach a point where it becomes impossible to tell the real from the fake, the situation will become highly problematic.
This can also mean an extended duration of many court cases. In Poland, where I live, the wait times for those are already notoriously long. Countries with similar problems can only expect things to get worse.
5. Entertainment
Recently, Justin Roiland, who voices the titular characters of the TV series ‘Rick and Morty’, was removed from the show due to domestic violence allegations. The creators have announced that they plan to continue the show, which has left many wondering how they are going to solve the problem of not having the original voices of the story anymore. This comes at a peculiar time, as we are on the cusp of achieving near-perfect voice deepfakes. Within the next few years, it may be entirely possible to solve issues like these by finding a replacement voice actor and using mimicking software to create a perfect imitation, assuming that it is legally permissible to do so.
Deepfakes are already being used in creative ways. Last year, the movie 'Fall' was supposed to be picked up for a theatrical release in the U.S. However, the film's frequent use of the F-word would have garnered an R rating, which could have negatively affected its potential box office performance. The producers couldn't afford reshoots, so they utilized deepfake technology and replaced the expletives with less offensive alternatives, successfully earning a PG-13 rating.
There have been attempts at de-aging actors for storytelling purposes, for example in 'The Irishman'; however, thus far, these attempts have not been entirely convincing. This, of course, will change within a few years at most.
Deepfakes open up a lot of possibilities for the entertainment industry.
6. A new use case for blockchain technology?
The blockchain is a decentralized, distributed ledger technology that enables the secure and transparent recording of digital data. So far, it has only been associated with cryptocurrency, but people have come up with many interesting ideas for potential applications, such as, for example, ensuring the legitimacy of votes in elections. Perhaps in the not-too-distant future, there will be a new, significantly important use case for blockchain - verifying the authenticity of audiovisual material.
7. Scams
Social engineering scams can already be pretty elaborate, but they're about to get a lot more sophisticated when we add the component of perfect impersonation. We should be much more vigilant in the coming years when a friend, family member, or co-worker asks us to wire money or even just provide some seemingly innocuous personal information over the phone or via video call.