The scary and exhilarating potential of synthetic media
Synthetic media is essentially artificially generated images, videos, audio, or text created with AI, through deep and machine learning. We have all come across some form of synthetic media, perhaps without even being aware, as they become more and more adopted for consumer and commercial purposes, and less distinguishable from reality.
For audio, we are familiar with Virtual Assistants such as Siri and Alexa that have the ability to synthesize speech and audio, chatbots which use natural language to synthesise interactive text conversations, and music generation through playing virtual instrument notes or generating waveforms that recreate instruments or human voice, such a platform launched by Google DeepMind called WaveNet.
Generating and editing 2D and 3D media is seen in animations, gaming and most popularly, images and videos. Using autoencoders and GANs, deep learning technology has created a new type of infamous synthetic media called Deepfakes – face swaps and facial re-enactments, which, despite their positive capabilities, have gained attention for not-so-good-intentioned uses: fake news, political polarisation and financial fraud. On the flip side, the technology has also cleverly been used for entertainment purposes on social media platforms.
One of the more exciting applications of synthetic media can be seen in synthetic reality. Synthetic reality is a platform for virtual avatars, influencers, and 'fake' humans. One popular example is Lil Miquela, the most well-known virtual influencer, created by Brud, with 2.8m followers on Instagram, who looks and acts so human she leaves people wondering if she is real or fake. Her influence goes as far as modelling for fashion brands and appearing in videos with real celebrities.
The implications are vast.
What synthetic media is seeing is the reduction of barriers when it comes to content creation. Time, costs, and physical equipment requirements are significantly reduced, minimising the process of idea to content and empowering creators with superior tools for communication and storytelling. This is a momentous push in the direction of content production democracy, with creativity no longer inhibited by lack of means.
The commercial applications are wide. Synthetic media allows businesses to meet personalisation demand in the content created for end-users, for example seamless translation of video content, or tailored scripts for individuals. Companies can enhance their product offerings, such as audio books improving humanness of their voices in the tone, emotion, speed, and pronunciation. Highly specific and cheap content can be easily and quickly produced to meet business needs, and predictions show that the COVID-19 pandemic will only accelerate the adoption. In an interesting concept we came across, AI’s combined capabilities may one day learn your digital behaviour and identity to play content made for an audience of exactly one: you. Think Black Mirror: Bandersnatch, but much more complex.
And where the production of a virtual avatar is possible, will this technology be able to allow humanity to archive themselves – digital immortality?
The controversy, predictably, stems from the blur in the lines of reality. As the technology behind the synthesis becomes more and more powerful, distinguishing real from fake creates a huge potential for misuse, misinformation, and malevolent agenda. And the lack of physical entities will create confusion in licensing and ownership for synthetic content, calling for a change in laws and regulation.
The research and development
The synthetic landscape has seen a huge number of start-ups joining the game to develop the potential uses of AI-generated media. As with most technological disruptions, those who fail to adopt often get left behind, which raises the question: how will you embrace synthetic media to advance your business in the new age of media?
Out of these start-ups, we looked at Synthesia, who develop technology that grants users the ability to internationalise and personalise video content with implications for advertising agencies, marketing campaigns, corporate communications, and e-learning. The company allows you to use their own avatars or your own, with synthetic or real voices to translate scripts into video content of up to 38 different languages, and the ability to create personalised videos at scale.
One direction of future research and development in this field is creating safeguards to regulate the technology – minimising harmful use of AI and building ethical standards into the new, and seemingly, limitless technology.
And here's some final food for thought:
What are your thoughts on where synthetic media can take us?