Editor’s note: This story was updated Aug. 13 to correct the spelling of AI startup Synthesia.
The past few years have seen artificial intelligence reach a tipping point, fueled by record levels of investment and the growing availability of training data. Yet what we see today is still only the tip of the iceberg when it comes to AI’s creative potential.
Various incarnations of synthetic media — content generated or manipulated by AI, often through machine learning and deep learning — have already started to be adopted for commercial purposes. These include virtual assistants, fashion models, and chatbots that synthesize text and speech, and we’ve also witnessed a wave of virtual beings, including computer-generated Instagram influencer Lil Miquela, KFC’s virtual Colonel Sanders, and Shudu, the first digital supermodel.
However, synthetic media has also become synonymous with deepfakes — videos or images where the person depicted has been replaced with another person’s likeness.
Untapped opportunities
As a result, the conversation has been distracted from some of the more exciting opportunities that synthetic media holds. Synthetic media will transform the way we produce and consume media. For the most part, it looks like AI will actually democratize creativity, rather than replace human content creators, and allow for greater experimentation.
Individuals will be able to produce high-quality content on minimal budget. Potential use cases include being able to go back and make changes to lines of dialogue spoken in a video or podcast simply by editing a text script, such as offered by the podcast tech startup Descript. Going a step further, AI startup Synthesia allows companies to create entire videos in multiple languages for purposes like staff training, simply by choosing from a selection of presenters and entering a script.
The COVID-19 pandemic will only accelerate the adoption of synthetic media. Restrictions on activities like video shoots may well become commonplace if we experience a second wave, giving way to the huge commercial opportunity for software that allows people to create deepfake-style videos. Additionally, synthetic media presents a way for influencers to scale their personal brand, with the ability to create digital twins that can represent them in films and commercials, etc.
AI-generated avatars also have some exciting potential use cases beyond the world of celebrity. The recent HBO documentary Welcome to Chechnya, which explores the persecution of LGBTQ people in Russia, used deepfake technology to protect the identity of interviewees by overlaying their faces with those of actors. Digital avatars also hold promise as a means of reducing bias and discrimination, such as in the context of recruitment.
Above: The current breakdown of synthetic media companies by type of media produced. Source: Samsung NEXT.
Above: Where synthetic media companies are based. Note that the actual number for China could be higher; information wasn’t available for some China-based startups. Source: Samsung NEXT.
Above: Founding years for the current slew of synthetic media companies. Source: Samsung NEXT.
Above: The current landscape.
How can we move forward?
In recent research into the evolution of the synthetic media landscape, my team has observed a huge uptick in the number of startups being founded in the past five years, with speech and voice synthesis and avatar synthesis representing the two largest sectors. Yet, unfortunately, many of these creative new use cases remain almost entirely overshadowed by concerns about deepfakes and misinformation in the discourse around synthetic media. As a result, we are failing to realize the disruptive potential of these technologies, which look set to usher in a new age of media.
These ethical questions are perhaps the biggest obstacle to the progression of synthetic media today. Every company working on synthetic media needs to take the time to consider an ethical code of conduct that they stand by — this includes acquiring the explicit consent of any party whose image or voice they use, and putting in place a strict internal screening process before releasing content to the public.
Industry-wide standards based around consent and misinformation will need to be implemented in a way that doesn’t stifle innovation, which will take time and require ongoing conversations at a national and international level. In the meantime, many synthetic media companies are taking matters into their own hands and building strict ethical codes into their products and partnerships. These startups are unlocking the potential of this technology now and enabling people to take creative expression to a new level in a way that’s mutually beneficial. It’s time to start telling the positive stories.
Iskender Dirik is Managing Director at Samsung NEXT Europe.
