Why we have reason to worry about this Deepfake application


Deepfake is a method of audiovisual content manipulation using machine learning. FILE PHOTO | NMG

You have to admit that humour goes a long way in relieving stress in what can be a very taxing daily existence. As technology continues to improve, creators are latching on this desire and need for light moments by building tools, services, platforms and apps that are benign at first look but whose use of technology opens a Pandora box.

Deepfake is a method of audiovisual content manipulation using machine learning. A recent rise to the global headlines is by the Chinese app – Zao, which was released on the iOS App Store. The app does one thing but does it well. It allows for users to convincingly superimpose their likeness on a character of their choosing, often a celebrity and thereafter share the content across the interwebs.

The concerns around deepfakes straddle the political, business and social existence. On the geopolitical front, it is not hard to imagine the use of the technology to generate content that can drive nations to war through a well-placed slur or outright executive order by a sitting president or in the context of a local by-election for a contender to be captured making certain false statements that leads to a drop in ratings or defeat should they have been a frontrunner.

In business circles, coupled with other social engineering attacks such as phishing, using speech synthesis will see an increase in corporate espionage and digital heists that will be near fool-proof and perpetrators will aim for bigger paydays under this new cloaks of invisibility. Socially, riding off the fact that most of us cannot stand any form of public embarrassment, especially on subject matters that are best discussed in silence or outrightly abhorred, we will see a rise in personal blackmail with high chances of ransom demands in cryptocurrency. I am sure you can think of many other possible scenarios where this technology can be put to detrimental use.

What stupefies me though, is the fact that the fun and riskier habit-forming applications of these technologies that offer nothing more than a temporary shot of dopamine, go viral with millions of users playing down the issue of potential abuse, terming their usage as harmless. All the while, truly ‘transformative’ - pun intended, use cases always seem to trigger some form or other of intense pushback in the same regulatory – or lack of, environments.

Which way it will sway beyond the experimental and fun remains to be seen.