Open Government, Civic Technology, Artificial Intelligence and Information Retrieval

Civic Technology Community Group,


Hello. I created and shared a summary of ideas with some other open government and civic technology individuals and forums (see the postscript).


These ideas can make a difference, describing exciting possibilities with respect to products and services for independent journalists, journalism organizations, nonprofits, government watchdog organizations, and others intending to gather, analyze, and disseminate public-sector content and data to their audiences.



Best regards,

Adam Sobieski



P.S.:


Introduction
The intersections of open government, civic technology, artificial intelligence, and information retrieval present many exciting opportunities for innovation.
For instance, multiple push-based information-retrieval services could spider and/or receive pings from public-sector websites when new multimedia content (e.g., documents, reports, transcripts, minutes, excerpts therefrom, and described datasets) arrived or when existing content was modified. Content and updates could then be automatically routed and recommended to interested citizens, journalists, nonprofits, and government watchdog organizations.

Pull-based Information Retrieval
Examples of “pull-based” information retrieval technologies include search engines and question-answering systems.

Push-based Information Retrieval
Examples of “push-based” information retrieval technologies include change detection and notification systems, intelligent content-routing systems, and recommender systems.
In theory, new artificial-intelligence technologies, e.g., foundation models, could be of use for intelligently routing content to interested users or groups of users based upon their models, stereotypes, roles, or responsibilities.
Explainable intelligent content-routing system nodes might utilize foundation models to answer natural-language questions, potentially executing dialogical workflows, about multimedia content. Such questions might resemble: “Would a journalist interested in climate, the environment, or environmental policy find this document to be interesting, and why?”
Explanations for routing and recommending content could accompany items through intelligent content-routing systems in the form of metadata. Such metadata could be utilized by subsequent content-routing system nodes, by users’ virtual-assistant agents, and returned when providing detailed feedback to information-retrieval services.

Social Information Retrieval
People could also share discovered content with one another via communication channels, social information-retrieval services, or by publishing articles or stories to their audiences.

Human-computer Interaction
With respect to user interfaces and user experiences, pertinent technologies to consider include syndication feeds, computer-generated digests, and agentic conversational user interfaces.
Users’ virtual-assistant agents could, in a configurable manner, perform attention-management on their behalf – coordinating with their natural-language-described tasks and determining whether and when to interrupt these tasks with alerts or notifications about routed and recommended content.

Machine Learning
Users could express their dynamic interests to their virtual-assistant agents and their push-based information-retrieval services, e.g., in receiving more content like items discovered through either pull-based, push-based, or social information-retrieval technologies.
Explicit and implicit user feedback with respect to routed and recommended content, thumbs up or thumbs down buttons or otherwise detecting usage, can enhance intelligent content-routing and recommender systems.

Usage Analytics Dashboards
Users could be provided with unified dashboards to view their feedback histories and other usage analytics with respect to multiple competing push-based information-retrieval services. This would be useful, for example, when customers wanted to conduct cost-benefit analyses with respect to multiple competing service subscriptions.

Standards and Recommendations
Existing and new standards and recommendations could be of use with respect to: (1) artificial-intelligence-generated metadata accompanying items routed and recommended through intelligent content-routing systems, (2) combinable streams of content items from multiple push-based information-retrieval services, (3) feedback from users to their multiple information-retrieval services, and (4) telemetry, logging, and other statistics collection, including to enable unified usage analytics dashboards for multiple competing information-retrieval services and subscriptions.