GLS Conference

Deepfakes, a product of artificial intelligence and software applications used to create
convincing falsified audiovisual content, have entered the international lexicon with
unsettling outcomes. Using deep learning algorithms, “deepfake” videos (or colloquially,
just “deepfakes”) typically substitute one person’s visual and acoustic likeness for
another, presenting viewers with compelling videos of individuals doing and saying
things they never did or said (Vaccari et al., 2020). Although verifying online content and
imagery is not a new phenomenon—in fact, the entire field of image forensics exists for
this purpose—academic research exploring this topic is only recently emerging. With
the assistance of sophisticated software programs,’ very exacting algorithms can create
high-quality videos and manipulated audio of individuals without consent. Deepfakes
are potentially the gateway to dangerous forms of crime, leading to the spread of
serious misinformation, often with devastating consequences. The detection rate of this
digital emergence proliferates exponentially, and the sourcing is challenging to verify,
causing alarms. It only takes one example to cause significant harm. Examples of this
information warfare are deepfakes that include identity theft, discrediting public figures
and celebrities, cyberbullying, blackmail, threats to national security, personal privacy,
intensifying pornography and sexual exploitation, cybersecurity, and baiting hate crimes,
abusing social media platforms, and manipulating metadata. These case uses create
pervasive and unaccountable misinformation that is often difficult to discern. This
presentation will review the state of research on deep fakes and their potential
implications on how users assess trust, possible afflictions on memory, and the future of
policies.