Post Authored by Melissa Getman

The internet is a wild and fast-paced world, which accelerates with every advance in technology. One such advancement in recent years has been the large influx of fabricated pictures, videos, and voice-overs involving “deepfakes”, or human image and voice synthesis generated through artificial intelligence (AI). Seemingly harmless deepfake content – like Saturday Night Live’s Bill Hader in conversation with David Letterman over an interaction with Tom Cruise, or the flood of Nicholas Cage’s face put on pretty much everything – have gone viral, but what’s the powerful technology fueling all of this content?

Generative Adversarial Networks (GANs) are the masterminds behind the operation, and have been used in industries such as the entertainment realm for years, giving producers and those alike the ability to create their artistry. Until recently, this technology was only accessible by the few, making it a low-grade threat to areas of our social fabric such as journalism, diplomacy, public safety, and national security. As accessibility skyrockets for this type of technology, the increase of these threats has followed suit, making people warry of the nefarious uses that could result from such ease of access. While the prospect of legal ramifications for these potentially catastrophic privacy infringements are readily apparent, the uses for such technology can also serve a much more positive endgame.

Lyrebird, a Canadian company that has made this type of AI available to the general public, has partnered with the Amyotrophic Lateral Sclerosis (ALS) Association with those goals in mind. Project Revoice, as the partnering companies call it, gives individuals who suffer from the degenerative disease the chance to preserve their voice for use after ALS claims their ability to speak. Imagine a loved one, suffering from this very common symptom, getting the opportunity to reclaim the ability to participate in conversation – a luxury most of us don’t think twice about having, and one that would understandably be devastating to lose. This has been the reality for Pat Quinn, co-founder of the infamous ALS Ice Bucket Challenge, as Lyrebird used this revolutionary AI to clone Quinn’s voice from various recordings of interviews Quinn had conducted surrounding the awareness challenge. Since Quinn’s success with the Lyrebird software, the company has helped others in the ALS community to reclaim the power of their voice.

As with most technological advances, with the good comes the bad. In the context of deepfakes, this means ease of access to the technology for the detriment, rather than benefit, of others. This technology makes two of the most trusted human senses – sight and sound – untrustworthy, and in a world where our need for authentication is continuously on the rise, deepfakes pose a real threat to the validity of many of the systems our society holds in high stature.

So, what’s being done about it? A few legislative pieces have been proposed that aim to deter individuals from knowingly creating or sharing this type of content, but the problem many of these drafters are facing is the adverse effects of over-regulation. If regulated too loosely, the deterrence desired is rarely achieved. If regulated too narrowly, this type of censorship could discourage the facilitation of organized speech. Stronger, more apt authentication technology is also on the rise to combat the surge of falsified information being generated with deepfakes, aiming to curb the accessibility concerns of individuals. It will likely invoke the balancing act that often comes within the legal world: the interests of the individual, the interests of technological advancement, and the interests of security.

References:

https://www.forbes.com/sites/kasiaborowska/2019/02/15/ai-gets-creative-thanks-to-gans-innovations/#fc36f5f708ca

http://www.alsa.org/news/media/press-releases/project-revoice-041218.html

https://slate.com/technology/2019/07/congress-deepfake-regulation-230-2020.html

https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201920200AB730

About the Author:

Melissa Getman is currently a rising 2L at DePaul University, College of Law, where she is involved in the Business and Commercial Law Journal, Phi Alpha Delta, the DePaul Law Advocacy Program, and the Business Law Society. Within the legal community, Melissa works as a law clerk at Stephan Zouras, LLP, which specializes in class and collective actions in the employment law realm. Melissa is also an active member of the Women’s Bar Association of Illinois, where she volunteers for various events such as the Judicial Reception, Annual Dinner, and Race Judicata. Melissa joined the Chicago community in 2018 as part of her move from South Florida, where she obtained a B.S. in Finance from Florida Gulf Coast University.