Man-machine Dialogues Grounded in Public-sector Meeting Transcripts

[Adam Sobieski]

US Open Government,

Hello. I am pleased to share some ideas pertaining to man-machine dialogues grounded in public-sector meeting transcripts. Citizens and journalists may soon be able to ask questions about and engage in dialogues with artificial-intelligence systems about public-sector meetings, the meetings of their city, county, state, and federal governments and public-sector organizations.

Here is a hyperlink to a blog article discussing these technical topics in greater detail: https://research.google/blog/efficient-data-generation-for-source-grounded-information-seeking-dialogs-a-use-case-for-meeting-transcripts/ .

Here is a quote from that blog article:

"Meeting recordings have helped people worldwide catch missed meetings, focus instead of taking notes during calls, and review information. But recordings can also take a lot of time to review. One solution to enable efficient navigation of recordings would be an agent that supports natural language conversations with meeting recordings, so that users can catch up on meetings they have missed. This could manifest as a source-grounded information-seeking dialog task where the agent would allow users to efficiently navigate the given knowledge source and extract information of interest. In this conversational setting, a user would interact with an agent over multiple rounds of queries and responses regarding a source text. The input to the agent model would include the source text, dialog history, and the current user query, and its output should be a response to the query and a set of attributions (text spans from the source document that support the response)."

In addition to enabling citizens and journalists to be able to manually engage in dialogues about one or more public-sector meetings' transcripts, search-automation scenarios are possible. Users' natural-language queries, questions, and "dialogical procedures" could be stored and applied to public-sector meetings' transcripts as they arrived. Users could, then, receive alerts, e.g., by email, as updates pertinent to their stored searches occurred.

I am excited about these ideas in the intersection of open government and artificial intelligence and wanted to share them with the group. Thank you.


Best regards,
Adam Sobieski