AI-Generated Characters: Putting Deepfakes to Good Use

Hybrid Workshop on April 30, 2022 at  ACM CHI 2022

AI-generated characters (more infamously known as ”deepfakes”) are a form of digital media generated by deep neural networks, which feature the human appearance and appear authentic to human beings. Such realistic renderings are made possible through advances in generative machine learning for the hyper-realistic synthesis, recreation, and modification of prose, images, audio, and video data. While the technology is perhaps most widely associated with media manipulation and misinformation, it is increasingly used for positive applications as well. For instance, it is integrated into industries ranging from entertainment, such as movie studios and games, to humanitarian efforts and educational tools. Given the growing interest in AI-generated characters across industries, we see a potential for future positive applications in a variety of areas including learning, privacy, communication, art, and therapy. This workshop will bring together researchers in HCI, AI, and related fields to explore positive applications and design considerations for AI-generated characters and related forms of synthetic media.

Suggested topics / areas:

• AI-generated media & AI-generated characters
• Case studies of the use of characters in HCI
• Virtual human creation & synthesis pipeline
• Learning with virtual instructors
• Health and wellbeing applications
• Generative Media Synthesis
• Human-AI collaboration
• Authenticity, Style and Improvisation
• AI Embodiment: Can the Mirror Lie?
• Design pitfalls and guidelines
• Ethics of AI-generated characters and their use in HCI

Call for Participation

We invite researchers from both academia and industry to submit a short position paper on the theme of AI generated characters and HCI. We will evaluate submissions on fit, ability to stimulate discussion, and contribution to the future of HCI. The workshop's website includes examples of past work in the area to help inspire and inform position papers. Papers should be a maximum of 4 pages, and should be submitted in the CHI Extended Abstracts format. The submission deadline is March 5th, 2022.

At least one author of each accepted position paper must attend the workshop and all participants must register for at least one day of the conference. We will host accepted papers on the workshop's website for participants and others to review prior to the meeting. Submission can be accessed through the website: http://deepfakes.media.mit.edu/.

Organizers

• Valdemar Danry, MIT Media Lab
• Joanne Leong, MIT Media Lab
• Pat Pataranutaporn, MIT Media Lab
• Pulkit Tandon, Stanford University
• Yimeng Liu, University of California, Santa Barbara
• Roy Shilkrot, MIT Media Lab
• Parinya Punpongsanon, Osaka University
• Tsachy Weissman, Stanford University
• Pattie Maes, MIT Media Lab
• Misha Sra, University of California, Santa Barbara