IGF 2020 WS #211 Collective human rights approach to deepfake applications

Subtheme

Organizer 1: Nardine Alnemr, Gulf Centre for Human Rights
Organizer 2: Khalid Ibrahim, Gulf Centre for Human Rights

Speaker 1: Khalid Ibrahim, Civil Society, Asia-Pacific Group
Speaker 2: Sam Gregory, Civil Society, Western European and Others Group (WEOG)
Speaker 3: Alex Comninos, Civil Society, African Group

Moderator

Khalid Ibrahim, Civil Society, Asia-Pacific Group

Online Moderator

Nardine Alnemr, Civil Society, Asia-Pacific Group

Rapporteur

Nardine Alnemr, Civil Society, Asia-Pacific Group

Format

Break-out Group Discussions - Flexible Seating - 90 Min

Policy Question(s)

How can deepfake development observe human rights? How can human rights principles for AI and deepfake be made accessible to all stakeholders? How to engage the different identities, interests and preferences of stakeholders in universal guiding principles for deepfake?

Identifying positive and malicious uses of deepfake by different actors. Devising accessible human rights principles for deepfake. Advising on principle for best practice in governing deepfake.

SDGs

GOAL 10: Reduced Inequalities
GOAL 16: Peace, Justice and Strong Institutions

Description:

Proliferating interest in harnessing artificial intelligence and automation in deepfake technology necessitates a human rights approach to discern the dual-use of deepfake. The session is at introductory level to make the conversation about deepfake accessible. The principles to be developed in this session too are to be make accessible for audiences of different technical skills level and disciplinary backgrounds. Workshop speakers will present three vignettes on the development and applications of deepfake content: DeepNude which ‘undresses’ women, deepfake on social media to reflect on content regulation by platforms, and OpenAI’s text-producing algorithms. After presenting the vignette, participants are asked to examine possible creative and malicious uses of the technology and discern possible universal principles to observe. The session aims to provoke insight from different perspectives in developing these principles: businesses, developers, users, and targets/subjects of deepfake technology. In light of this, the workshop will probe into implications of absent guidelines for deepfake in situations such as: documenting conflict violations, attacking freedom of speech, and targeting members of vulnerable communities. For this purpose, highlighting the intersection between identity, stakeholder group and regional contexts is important to capture the challenges in leaving deepfake unattested by stakeholders. Speakers in this session will join the participants in developing the principles to allow for exchange regarding the vignettes, as well as ideas geared towards developing the principles.

Expected Outcomes

The direct outcome of this session is producing a 5-10 principles guiding businesses, developers, and users of deepfake applications. By the end of the session, participants will be invited to join the organisers’ mailing list if they are willing to sign the principles for endorsement before publication. Indirectly, the session aims to initiate a conversation about the role of different stakeholders in ensuring that human rights are observed in technology development and application. Particularly, identify when and what collective effort can do to safeguard the creative uses and combat malicious employment of deepfake. The potential of this conversation to be expanded in other spaces post-IGF.

Hosting the session online for attendees.

Relevance to Internet Governance: Safety, security and inclusion of stakeholders is one of the critical aspects of open and free internet. To achieve this requires using spaces such as IGF to discuss the principles to be sought in navigating the technologies which affect our online interactions and the internet. Therefore, this workshop in itself resembles a governance process in which stakeholders come together to think about advancing the objective of safe, secure and inclusive internet and means for enforcing these principles and objectives in light of deepfake technology.

Relevance to Theme: The questions raised in this workshop include who is included when thinking about human rights and emerging technologies? And to what extend do identities, interests and preferences influences by personal, professional and spatial experiences are articulated in contributing to internet governance? This workshop fits in the thematic track of inclusion because while entertaining these questions, it aims to illustrate through the workshop and its outcome that diversity and inclusion are important values in thinking about internet governance, discussing internet governance, and the creation of norms of internet governance. In effect, inclusion can be achieved even when not present on the table by thinking about principles that are accessible and easy-to-understand by stakeholders who are of limited knowledge of the debate.

Online Participation

 

Usage of IGF Official Tool. Additional Tools proposed: Zoom, to stream the discussion and organise questions and follow of the conversation