So what are deepfakes?
Somewhere out there on the internet, there is a video of Barack Obama giving a public service announcement warning of the dangers of fake videos and calling Donald Trump a very nasty word in the process. There is also one in which Mark Zuckerberg supposedly speaks frankly about how he basically wants to “own you” through Facebook. If you have not yet seen these videos you really should have a look. Because these are prime examples of deepfakes. In Layman’s terms, they are an advanced form of photoshopping. They use a type of Artificial Intelligence called deep learning to create very realistic footage of fake events. Thus the name deepfake.
But it’s not just videos. Deepfake technology can also create entirely fake photos from scratch that will have you convinced you’re looking at the real thing. LinkedIn, for example, has quite a few profiles with those. But it doesn’t stop there, audio can be deepfaked too — the so-called voice-cloning deepfakes. Some public figures and higher-ups in prestigious firms have already been scammed by blackmailers using these. The blackmailers threatened to leak to the public voice recordings of these figures saying damaging things they never said in a voice that is damningly identical to their own unless they paid up.
And deepfakes, in all their variations, can come from anywhere. Pretty much anyone with access to the technology is getting in on the action of making them, from academic and industrial researchers to amateur enthusiasts to porn producers. Governments are dabbling in it too, albeit mostly to discredit extremist groups.
While deepfakes can certainly cause trouble on an individual or community level, their potential to spark major international incidents is still, thankfully, limited. A deepfake satellite image of troops gathering on a border between two countries on the brink of war, for example, will not push any decision-makers to regrettable actions. Most nations already have their own security imaging systems down pat.
But their potential for wreaking havoc in the lives of unsuspecting people remains problematic. Just think of child custody battles. Someone could lose the right to see their child as a result of deepfaked events that were entered and accepted as evidence in court. And here’s another scenario: people could lose a hefty chunk of their savings because of a deepfaked video call on WhatsApp in which someone who they thought was their mother or brother asked them to make a wire transfer to an unknown bank account.
They also present opportunities for public deception — many public officials can use this as a tactic to not be held accountable. If they’re caught on camera doing or saying something they shouldn’t, they can simply cry deepfake and avoid public demands for resignation. Or at the very least cause mass confusion, which more often than not leads to mass amnesia.
So what can be done to stop this?
While this technology uses Artificial Intelligence to carry out its mischief, it is, ironically, AI itself that can be used to stop it — AI is already being used plentifully to spot fake videos. And tech companies are working on detection systems that will pinpoint fakes whenever they pop up. Also, blockchain technology may come to our rescue as well. This technology namely uses an online ledger system that keeps a record of videos, pictures, and audio. This record would be impenetrable so any attempt at foul play or manipulation of these footages would fall through the cracks and can always be checked and rectified.
Believe it or not, there is an upside
Despite overwhelming evidence to the contrary, deepfakes aren’t all bad. The voice-cloning deepfakes, for example, can also be used to give people their voice back when they lose it to disease. And as for the videos, they can also be used for entertainment. They’re already becoming an attraction at galleries and museums, where they bring historical figures to life. They can also be used to improve the dubbing for foreign-language films and resurrect dead actors to star in new Hollywood films. But as you can imagine that last one is quite controversial.
So all in all, as long as we keep in mind that technology can have a pretty and ugly side and keep putting safeguards in place to shield us from the effects of the ugliness, deepfakes should have minimal impact on the way we shape our reality.