IPTD-413
OpenAI Access (evaluation of public comments using Artificial Intelligence)
Description
No description
Problem Statement
The NJ Forest Service (NJFS) frequently takes on projects that require the processing of large amounts of stakeholder comments. Recent updates in Large Language Models (LLM’s) have shown to be successful in classifying and parsing this type of data. The NJFS would like to pilot and evaluate the use of OpenAI’s interface in using such models for the update to the NJ State Forest Action Plan and the Wharton State Forest Natural Resource Stewardship planning effort to be able to evaluate the use of these technologies for more effective public comment process across NJDEP.
Project Justification
An experienced team in NJFS recently completed an analysis of comments received in response to a series of surveys regarding the projects at Liberty State Park, at the request of the Assistant Commissioner’s office on an experimental basis. To accomplish this, NJFS used the OpenAI API because this interface has excellent documentation and a well-developed API. The technology showed a great deal of promise on that project by cutting completion time by more than half.
This project will use the OpenAI interface or similar to leverage LLM’s to process public and stakeholder comments and extract meaningful data improving NJDEP’s ability to respond to comments. A dedicated team of four specialists in NJFS that has experience with AI and LLM’s will use these models to analyze public comments and automate some of the process using scripts.
Deliverables include analyses of stakeholder comments, scripts and code for increasing flexibility for LLM use.
This project fits within the guidance of Joint Circular NO.: 23-OIT-007, STATE OF NEW JERSEY INTERIM GUIDANCE ON RESPONSIBLE USE OF GENERATIVE AI and Executive Order-346 regarding exploring the use of AI to improve government services.
This project will use the OpenAI interface or similar to leverage LLM’s to process public and stakeholder comments and extract meaningful data improving NJDEP’s ability to respond to comments. A dedicated team of four specialists in NJFS that has experience with AI and LLM’s will use these models to analyze public comments and automate some of the process using scripts.
Deliverables include analyses of stakeholder comments, scripts and code for increasing flexibility for LLM use.
This project fits within the guidance of Joint Circular NO.: 23-OIT-007, STATE OF NEW JERSEY INTERIM GUIDANCE ON RESPONSIBLE USE OF GENERATIVE AI and Executive Order-346 regarding exploring the use of AI to improve government services.
Estimated Transactions
Dependent on number of activities involving public comment related to
Components
Large Language Model (LLM)
Target Rollout Date
15 March 2024
Target Rollout Date Reason
03/15/2024 – Would provide adequate time for rollout of stakeholder comments, for a number of planning projects including. Use of LLM’s will help NJDEP to more effectively process and respond to stakeholder and public comments, improving public involvement in a number of projects
Activity
Show:
Asked Bing: i need to compare tools for summarizing large sets of comments, categorizing and quantifying their occurrence and strength of sentiment
Sure, here are some tools that you might find useful for each of your needs:
Summarizing Large Sets of Comments:
QuillBot: This tool can condense articles, papers, or documents down to the key points instantly1.
Scribbr: It offers a free summarizer tool that can generate unlimited summaries without paying a penny2.
Jasper: An AI summarizer tool that can create high-quality copy and summaries from texts3.
Categorizing Comments:
Doakio: It offers a comprehensive comparison of text classification tools4.
MonkeyLearn: This tool uses natural language processing (NLP) to understand the context behind social media posts, reviews, and feedback5.
G2: It provides a list of commenting systems that allow users to engage with posted content via comments or reactions6.
Quantifying Comment Occurrence:
MeasuringU: It offers a guide on how to code and analyze verbatim comments7.
Scribbr: It provides a guide on how to conduct content analysis to quantify the occurrence of certain words, phrases, subjects, or concepts in a set of historical or contemporary texts8.
Measuring Strength of Sentiment in Comments:
Sprout Social: It offers a list of sentiment analysis tools that use artificial intelligence (AI) to analyze textual data and pick up on the emotions people are expressing9.
HubSpot: It provides a list of sentiment analysis tools that analyze text conversations and evaluate the tone, intent, and emotion behind each message10.
Ascribe: It offers a feature within CX Inspector that provides a sentiment score from open-ended comments11.
These tools can help you summarize, categorize, quantify, and measure the sentiment of large sets of comments. Please note that the effectiveness of these tools can vary based on the specific requirements and the nature of the comments you are analyzing.
Project Planning Sheet sent to DoIT. This project fits the recent ask for DEP AC program areas to submit projects that wish to leverage Artificial Intelligence.