The Online Deliberation Platform is a video discussion platform for groups of 8-15 people. The platform is designed to facilitate a structured and equitable conversation with better opportunity for participants to speak up. It is developed by the Stanford Crowdsourced Democracy Team in collaboration with the Deliberative Democracy Lab.
The system assigns speaking turns, manages a speaker queue and makes sure people don’t talk for too long.
The convener defines a structured agenda with topics and a balanced set of arguments for each topic. The system ensures that the conversation finishes on time.
When the system detects toxic behavior, or when a participant is reported, the automated moderator asks the group to validate this and takes action accordingly.
The platform enables more equitable participation from all participants using a speaker queue and will nudge participants that have not spoken much.
No need to download or install anything. The website works directly in web browsers. We currently support Chrome and Firefox on desktop and Android devices.
The analytics tools let the convener monitor and visualize the conversations in real time. This includes live transcription, speaking time distribution, and audio recording.
We support up to 100 simultaneous conversation groups with up to 15 people in each group.
The platform currently supports English, Chinese, Cantonese, French, Japanese, Norwegian and Spanish. More support can be added on request.
The platform is under active development. We support features such as closed captioning, screen sharing, question development, and agenda building tools.
The convener can choose from a range of configurations that suit their needs. For example, you can choose to only support audio, set the speaking times, and so on.
The Online Deliberation Platform is a video discussion platform for groups of 8-15 people. The platform is designed to facilitate a structured and equitable conversation with better opportunity for participants to speak up. It is currently in limited alpha mode: there is limited capacity available to test/use the platform.
The platforms runs in Chrome or Firefox without installing any additional software on desktop platforms and Android. The convener defines a structured agenda (e.g., "Electoral Reform") with topics (e.g., "Move all primary elections to after Labor Day") and a balanced set of pro's and con's for each topic.
After a brief set of instructions, the platform invites participants to introduce themselves, and walks them through the agenda, displaying for each topic the set of arguments. Participants take turns speaking for up to 45 seconds, and join a queue when someone else is speaking. The platform manages both speaking time and queuing, allowing brief interruptions. When the group thinks all arguments have been discussed, they can propose and vote to move on to the next topic. When all topics are discussed, the group enters the optional concluding phase, where they collaboratively develop open questions, which are discussed, edited and voted on as the outcome of the conversation.
Because of its dynamics, the platform enables more equitable participation from all participants, and will nudge participants that have not spoken much. When the system detects toxic behavior, or when a participant is reported, the automated moderator asks the group to validate this and takes action accordingly. The convener can monitor the conversation live, and record the session’s audio and interactions.
The platform is under active development, and has recently been successfully deployed in Canada, Chile, Hong Kong, Japan, and at Stanford with 15 parallel groups each. We have supported deliberations with up to 1250 unique participants. So far, we have hosted over 1450 hours of group discussion with 7000 unique participants.
You can read our demonstration paper at HCOMP and our news coverage here.
This work is supported by the Stanford Institute for Human-Centered Artificial Intelligence through a seed grant, the Stanford Research Institute and previously by the David and Helen Gurley Brown Institute for Media Innovation through a Magic Grant and the National Science Foundation.
Our platform has been used in several deliberations in various contexts.
Some examples include:
If you are interested in hosting a small-group conversation on the platform or want to learn more about it, please email us at firstname.lastname@example.org
We would like to thank our past members, who have made significant contributions to the platform: Nikhil Garg, Xinyu Hu, Daniel Kharitonov, Isa Liang, Arafat Mohammed, Liubov Nikolenko, Shouzhuo Sun, and Sravya Yandamuri.