Deepfakes and synthetic media: The new age of disinformation is growing

Tuesday, 21 May 2019

Artist Jessica Savage Broer photoshopped this image to include President Donald Trump. She said she wanted to make the point that people need to use critical thinking skills.

Amidst the threats identified in a new report of the digital threats to democracy are two challenges with sinister sounding names - deepfakes and synthetic media. But what are these emerging risks and why should we care about them?

The issue is explained in a Law Foundation-backed study by Tom Barraclough and Curtis Barnes of research company Brainbox, a think tank that looks at the intersection of law, policy and emerging technologies.

The authors called their study: Perception Inception, Preparing for deepfakes and the synthetic media of tomorrow.

There's already fake news, and plenty of different ideas about what is and isn't fake, but we're speeding toward a time when it's going to become increasingly difficult to know whether you can believe what you're seeing.

Aristomenis Tsirbas, a Los Angeles-based director, made a computer-generated video of a plane doing a 360 that ended up being repurposed as part of a fake news report.

* Fake-porn videos are being weaponised to harass and humiliate women

* I fell for Facebook fake news. Here's why millions of you did, too**

WHAT IS SYNTHETIC MEDIA?

In their report, Barraclough and Curtis said they used the description to essentially mean audiovisual information in digital form. Often it was a composite of multiple pieces of information synthesised to produce something substantially new.

Even modern digital cameras were synthetic media technologies, they said. The photographs they produced were composites constructed from data that may be collected by multiple sensors. Various manipulations could be carried out intended to enhance the information. For example, making it more visually pleasing by removing such things as 'red eyes'.

Wonder Woman star Gal Gadot has been a victim of face-swapping fake-porn videos.

The recent picture of a supermassive blackhole was an excellent, even if unusual, analogy to explain synthetic media. The picture was a composite synthesised from a planet-scale array of eight different telescopic sensors around the globe, acting in synchronicity.

Not only that, the sensors used were radio telescopes detecting radio waves imperceptible to the human eye, converting electromagnetic information into digital data that was then used to generate a visual image.

WHAT ABOUT DEEPFAKES?

Deepfakes are one example within the wider family of emerging audiovisual synthetic media technologies. The title deepfake is often used for video and audio synthesised from existing digital data using deep learning neural network models - a kind of machine learning artificial intelligence technique.

Deepfakes can be used to generate high-quality visual and audio representations of things that never happened. 'In New Zealand, our experience is that most people are yet to encounter these sort of audiovisual artefacts – deepfake videos, synthetic speech, and so on – despite their increasing commercialisation,' the report said.

All it takes is a smartphone to perpetrate a deepfake - and the technology's getting better all the time.

'Few New Zealanders are likely to be aware that at present, from around five to 10 minutes of video or 20 to 30 minutes of audio, a skilled person using consumer-level computing technology could create relatively realistic representations of the prime minister engaged in entirely untrue behaviour, or saying totally fabricated things.

'With enough video or photographs, perhaps taken from a personal Facebook or Instagram account, a skilled person could even produce such misleading material of an everyday New Zealand citizen.'

The meaning of the word 'deepfake' was fluid and had expanded beyond its original meaning to become a catch-all for any form of audiovisual falsity or manipulation. It had been used alongside the phrase 'information apocalypse'.

HOW BAD IS IT?

To give some indication of the implications, the report authors referred to an ABC News interview with Dr Matt Turek, head of the media forensics programme at the Defence Advanced Research Projects Agency, run by the US Department of Defence.

'The challenge is it goes to the heart of our trust in visual media,' Turek said. 'We're used to looking at an image or video and putting faith in it -- believing the content of the image or the video. And with the rise of the ability to easily manipulate those, that is going to put our faith in visual media in jeopardy.'

According to the ABC item, deepfakes began sparking widespread concern in 2017 when Reddit users started posting fake pornographic videos online. They primarily targeted actresses such as Wonder Woman star Gal Gadot, superimposing the superhero's face onto X-rated content.

'The early fakes were riddled with glitches, but as that technology continues to evolve, some worry they could become indistinguishable from the real deal – potentially swaying elections, triggering widespread panic, riots – or even a war,' ABC said.

'A lot of times there are some indicators that you can see, particularly if you are trained or used to looking at them. But it is going to get more and more challenging over time,' Turek said.

The challenge was that with new technologies, it was much easier than in the past for one person to create manipulated video or images. And it's much easier to circulate the manipulated content.

Then there's the 'liar's dividend', when it's not clear what's true, a dishonest person could thrive by saying something is fake when it's actually true.

OTHER THAN FAKE PORN, WHAT'S ANOTHER EXAMPLE?

False social media accounts trying to gain access to information, or spreading information. One case was the fabricated senior journalist 'Maisy Kinsley', who was supposedly working for Bloomberg. Kinsley had a fake LinkedIn profile with 195 connections, and a personal website that may have been written by an algorithm.

Kinsley's Twitter account followed some Tesla short sellers, at least one of whom reported the account tried to get personal information from him. It's thought possible the aim of the account was to spread information that could affect the value of Tesla stocks.

IS THE PROBLEM GETTING WORSE?

As the Brainbox report noted, people now are able to manipulate information and disseminate it to the world using just a smartphone.

'Moreover, the capacities of these technologies are rapidly improving. Common applications like Snapchat filters, augmented reality, and so-called 'face swapping' allow everyday people to create and share more and better manipulated audiovisual information: information that makes it look or sound like something happened when it did not,' the report said.

'It illustrates the way that these technologies are likely to become commercialised and converted to consumer products requiring little expertise to operate and having largely innocuous impacts. Many of these applications share the same or similar techniques as deepfake videos. These sorts of technologies are becoming subtly but extensively pervasive.'

WHAT IS IT GOOD FOR?

These technologies have many commercially valuable and beneficial applications, the report said.

'As creators New Zealand companies are world leaders in audiovisual effects technologies. The services of Weta Digital, for example, are highly sought after across the world and have been central to countless blockbuster films.

'Companies like Soul Machines sit at the forefront of applied computing in their field, pushing the boundaries of animation technologies,' the report said.

'New technologies allow for the replication of faces and voices, sometimes characterised as 'digital likenesses'. This is particularly relevant for persons who trade on their face, voice, or performance in some capacity, or the people who capture and distribute those performances in commercial markets.

'The advancing technologies of virtual human avatars have potential for use as substitutes (or supplements) to real human actors, as does the use of deep learning neural networks to enhance the flexibility of post-production techniques.'