DeepFake – Machine Learning Hero or Villain?

If you’ve not seen a DeepFake Video yet, or even heard of DeepFake Videos – you soon will. And that is despite Twitter and Reddit both banning them in the last few days. Technology drives our society forward – but yet, technology itself is passive. It’s how technology is USED by humanity that matters most.

It is said that many of the members of the Manhatten Project – the developers of the Atomic Bomb in the 1950s – were so caught up in the technology and the scientific breakthroughs that they were achieving that the wider questions of whether it SHOULD be done were largely pushed to one side. It was only after they saw the effects of their creation that they had second thoughts. Whether the creators of DeepFakes have the same soul-searching questions remains to be seen.

Deepfakes are a marvellous example of Machine Learning (ML) given a practical, obvious, easily-understood, real-life application. The fact that most of the current uses are somewhat “dodgy” is, however, a real concern.

It’s a well-known fact that the adult industry has historically been a leader in the technology field. It’s simple economics – the adult industry is so huge and worth so much that it can have such a huge influence on business and wider society. Kids and young adults today brought up on Netflix and Amazon Prime (and similar music-based streaming services) can be forgiven for finding it strange that we used to have physical representations of the entertainment that we bought. For anyone who thinks that Blu-Ray & DVD is so passé, the very existence of VHS And Betamax would be a complete mystery.

Betamax was the superior technology – but it had very stringent rules and regulations which hampered more ‘creative uses’ of videotape technology. VHS, despite not being as good, was more open and ‘liberal’. As such, the adult industry had no choice but to go down the route of VHS. The rest (so far as tape-based video media is concerned) is history: Betamax died out with little fanfare while VHS reaped all the spoils. The ‘mainstream’ publishers had no qualms whatsoever in following the lead that the adult industry set: even though they knew Betamax was the better product, they knew that VHS had become the better bet.

The wonderful thing about the democratisation of technology is that anyone can do whatever they have the ability to do. The terrible thing about the democratisation of technology is that anyone can do whatever they have the ability to do.

The technology landscape is different now – the power isn’t in the hands of a select few with $$$$ behind them. It’s up to whoever has the right computing power, the right skills, and the right motivation to produce whatever they like. The wonderful thing about the democratisation of technology is that anyone can do whatever they have the ability to do. The terrible thing about the democratisation of technology is that anyone can do whatever they have the ability to do.

A DeepFake essentially is a video that transposes one person’s head onto another person’s body by using Machine Learning techniques. Perhaps the most reasonable use of this technology that we can discuss here is to create spoofs of movie stars appearing in other movies. Anyone who knows their memes would know that Nicholas Cage would be a prime use case for this new technology. And so it proved.

Take this a step forward, the same techniques can be used in political satire. It has been said that the only true measure of a strong, liberal society is if satire is allowed free reign. In the UK, particularly, where political satire has been a staple on TV for almost as long as TV has been around, it might seem that questioning if satire is allowed or not is simply silly. But that’s to take for granted the society we live in. In other countries around the world, satire doesn’t have the same free hand. Using DeepFake technology for purely satirical purposes in the UK would perhaps be seen as clever and humorous – yet the same technology used in other countries could be seen as an act of treason or sedition – and therefore result in the consequences of such acts.

But let’s just imagine that satire is ok… but then let’s take it just a step further still. It is so easy to imagine DeepFake technology that obviously ridicules a politician to be subverted to try to portray a politician in a situation or a manner that misrepresents them completely. The Fake News controversy over the past few years centred largely on simple written content. This could surely be dwarfed by the effects of fake videos portraying people saying or doing things that they have not done, or would not even do, in reality.

And the above is just the topics that we can safely talk about here. The ‘creations’ that have made DeepFakes infamous over the last few months have been much less suitable for a serious technology site such as this to talk about.

But the point remains – DeepFakes are only possible because of the recent technology advances that have been developing over the last few decades. Whilst the software technology to produce DeepFakes has been made available freely, the hardware technology, and the individual skill to take full advantage of that software is still limited to those that can exploit it.

Without the right hardware and without the right skill, the results obtained are laughable. But a bit of skills, a bit of motivation, and the access to the right hardware, then almost anything is possible.

And that’s where The Cloud comes in. Fifteen years ago – even less – the best way to exploit the best consumer computer processing power would be to spend a huge outlay and buy the hardware yourself and sit it on the desk in front of you. Now, you don’t need the processing power sat on top of your desk – it can reside in the cloud at the end of a broadband connection. GPUs in the cloud are already an obvious use case of The Cloud – Cloud GPU use over the next few years is going to get bigger and bigger.

Alvin Feague of DeepFake.ML (DeepFake Machine Learning) is quite optimistic, however. Despite the less-than-tasteful origins of DeepFake Videos, he tells us that, “the [machine learning] technology behind them is sound,” and that rather than the, “slippery-slope” that some have described the current deluge of DeepFakes, the techology is a, “rocket ship about to take off!”.

“The Genie is out of the bottle,” Alvin continues. Although he doesn’t expect distasteful DeepFakes to disappear, he does expect that more mainstream uses will come to the fore: just as adult content made VHS, and the mainstream industry reclaimed it, Deepfake Videos of a more “questionable” nature won’t go away – but more palatable be brought more and more into the mainstream.

+ posts

CIF Presents TWF – Professor Sue Black


Related articles

How Businesses Should Tackle Big Data Challenges

In today's data-driven landscape, Big Data plays a pivotal...

UK IP Benefits and How to Get One

There are many reasons why you may get a...

Navigating the Landscape of AI Adoption in Business

In today's rapidly evolving technological landscape, the integration of...

Three Ways to Strengthen API Security

APIs (Application Programming Interfaces) are a critical driver of...

A Comprehensive Guide To The Cloud Native Database [2024]

Databases are crucial for storing and managing important information....

Subscribe to our Newsletter