IGF 2023 WS #455 Toward Responsible Generative AI: Promises, Pitfalls, Policy


Artificial Intelligence (AI) & Emerging Technologies
Chat GPT, Generative AI, and Machine Learning

Organizer 1: Rachel Adams, Research ICT Africa / Global Index on Responsible AI
Organizer 2: Fernando Perini, πŸ”’
Organizer 3: Matthew Smith, πŸ”’
Organizer 4: Shachee Doshi, USAID
Organizer 5: Yunwei Aaryn, Global Affairs Canada
Organizer 6: Schroeder Zara, πŸ”’

Speaker 1: Rachel Adams, Civil Society, African Group
Speaker 2: Fernando Perini, Government, Western European and Others Group (WEOG)
Speaker 3: Matthew Smith, Government, Western European and Others Group (WEOG)
Speaker 4: Shachee Doshi, Government, Western European and Others Group (WEOG)
Speaker 5: Yunwei Aaryn, Government, Western European and Others Group (WEOG)


Fernando Perini, Government, Western European and Others Group (WEOG)

Online Moderator

Schroeder Zara, Civil Society, African Group


Schroeder Zara, Civil Society, African Group


Break-out Group Discussions - 60 Min

Policy Question(s)

A. How can generative AI support international development and the realisation of the Sustainable Development Goals?
B. What risks to human rights and democratic values does the use of generative AI in developing contexts pose?
C. What kind of global governance arrangements are needed to safeguard the development and use of generative AI in developing regions?

What will participants gain from attending this session? Participants will build understanding around the future uses and implications of generative AI in developing contexts. They will also learn of work being planned by key donor organisations in this area, and opportunities to be involved. The participants will have the opportunity to comment on and validate the horizon scanning report on generative AI in international development being undertaken by D4D.net prior to its publication. Lastly, as the discussions of this session are fed back into various global and regional policy processes, participants will have the opportunity to participate in shaping the governance of this emerging technology.


Generative AI represents a major step-change in the advancement of AI technologies, moving from computer intelligence capable of detection and prediction, to intelligence able to create human-like digital material. This technology holds huge potential to aid in advancing the realisation of developmental goals; there are, too, huge risks to digital inclusion, human rights and democratic values. In this session, we will explore what global governance tools and research agendas are needed to better understand and safeguard the use of generative AI in international development. This will build on the work of the D4D Network of Global South research institutions in understanding the implications of generative AI in diverse contexts.
To provide some researched background on this topic, the session will begin with a brief presentation on the findings of horizon scanning research undertaken by the D4D Network to explore key trends, concerns and priority areas related to the potential role of generative AI. The presentation will be followed by a mentimeter activity to engage the onsite and online audience on their reactions to the key issues and opportunities identified in the research. While the mentimeter is running, a respondent will offer insights on the horizon scanning, outlining priority areas for international development and global policy. The onsite session chair will then introduce the session activity to workshop priorities for global governance and policy in this emerging area. 20 minute breakout groups will be facilitated with the online and onsite attendees, structured around a series of prompting questions. The session will close with a quick report back from the breakout groups and a short overview of the way forward and how participants can continue supporting and following the work.

Expected Outcomes

The immediate outcome of this session will be in the publication of the D4D.net horizon scanning report on the role of generative AI in international development. More broadly, this work and sessions is intended to support the identification of priority areas for further research, international development assistance funding, and regulatory decision-making. It will support in determining the research and funding agendas of the International Development Research Centre of Canada, Global Affairs Canada, USAID and the Swedish International Development Agency. In addition, we intend for this work to provide key evidence in support of global governance processes on generative AI, including that of the G7 group following the meeting in Hiroshima in May 2023, and the ongoing work of the Global Digital Compact.

Hybrid Format: Onsite and online attendees will be engaged at the beginning of the session through a mentimeter exercise, the results of which will be presented and discussed. A third of the session (20 mins) will be dedicated to break-out rooms with online and onsite participants to facilitate engagement and discussion about the topic of the session. Online participants will be split into groups of 10, and supported by a dedicated online facilitator from our team. A specific member of the team will be online and dedicated to supporting the participation of online participants.