Team 5- DelegAI

Track Name: GovTech

Challenge Statement: Challenge Statement Document

Members: @nurs & @sum & @Spiderman2099

Team Lead: @nurs

Challenge Statement:

Some delegates use their votes just by following up to delegates with a certain amount of voting power, how to fix this? Besides, how Arbitrum DAO forums can be used effectively to make discussions less time consuming and how to reach out to more people out of the DAO?

Constitutional / Non-Constitutional:

Constitutional

Video: video
Discord: [DelegAI - Google Slides]

Abstract:

Make a Snapshot and Tally extension called “DelegAI” which allows to highlight the key points and comments in the governance forum, for each proposal. In order to do this, the information needed will be obtained from Discourse using AI.

Motivation:

As DAOs are developing and unfolding into something more complex, governance forums are becoming harder to read and overwhelming. This could be an obstacle for many people for them to join Arbitrum DAO and take a part in governance in general. In order to compensate for that, delegAI offers the new users a new experience letting them have an oversight of the governance forum discussions and save time on deciding on their votes.

Rationale:

This AIP asks for 96.5$ of cost for a tool development on summarizing the key highlights from the governance forum and reflecting it on Snapshot and Tally for the delegates to see while deciding on their vote which will eventually let us achieve:
•⁠ ⁠more DAO engagement
•⁠ ⁠onboard new delegates to Arbitrum DAO
•⁠ ⁠more engagement on Arbitrum DAO governance forum
•⁠ ⁠better voting rates
•⁠ ⁠prevent any mistakenly led voting quorums

Steps to Implement:

The tool gets all comments on the proposal from the governance forums via the API.

For sentiment analysis, we use an DistilBERT, an open-source, pre-trained NLP model that works with high accuracy on comments from many previous forum discussions. This tool decides on the vote derived from the forum opinions and uses this information to automatically vote for the proposal. Then we use data analysis tools in order to provide visual results about the statistics for anyone who would like to review.

The voting process is carried out by a DAO contract.

Timeline:

Three months after proposal passing the DAO.

Overall Cost:

For the overall cost we will be requesting a cost to only cover expenses for the development of the tool. Which are:

•⁠ ⁠slack 7.25$ / 12.50$ monthly
•⁠ ⁠docker 24$ (per user/month)
•⁠ ⁠vercel 20$ (per user / month)
•⁠ ⁠github 40$ monthly

Making the total cost to be 96.5$ monthly.

3 Likes