ChatGPT, as a sophisticated Large Language Model, excels in understanding human language. Instead of merely seeking answers from ChatGPT, you can leverage its natural language understanding (NLU) abilities to analyze existing information. In this article, you’ll discover:
- How to have ChatGPT summarize lengthy texts
- How to extract key information from passages
Let’s dive in.
Overview
This article is structured into three main sections:
- Your AI Secretary
- Extracting Unstructured Information
- How to Use ChatGPT for NLU
Your AI Secretary
Imagine having a meeting transcript that captures discussions among various participants. While this transcript is useful for understanding how conclusions were drawn, converting it into clear meeting minutes makes the information much more accessible. ChatGPT can help you transform such transcripts effectively.
Example Transcript
Consider the following excerpt from a meeting transcript:
GOOD MORNING, EVERYBODY. I’M MICHAEL STERN, DELEGATE FOR COMPTROLLER BRAD LANDER AND CHAIR OF THE FINANCE COMMITTEE OF THE NEW YORK CITY TRANSITIONAL FINANCE AUTHORITY. AS A REMINDER, PLEASE KEEP YOUR FILMS MUTED UNLESS SPEAKING. I’LL NOW CALL TO ORDER A MEETING OF THE FINANCE COMMITTEE. RITA, COULD YOU PLEASE CALL THE ROLL?
…
THANK YOU, EVERYBODY. HAVE A GREAT WEEKEND.
HAVE A GREAT WEEKEND.
THANK YOU.
While this is a brief meeting, the transcript may cover several pages. You can ask ChatGPT to summarize it into concise meeting minutes using the following prompt:
“As an experienced secretary, please summarize the meeting transcript below into meeting minutes, including the participants, agenda, key decisions, and action items.”
ChatGPT’s Output Example:
Meeting Minutes
Meeting: Finance Committee Meeting
Date: [Insert Date]
Time: [Insert Time]
Location: [Insert Location]
Participants:
- Michael Stern (Delegate for Comptroller Brad Lander and Chair of the Finance Committee)
- David Womack (Representing Director Jacques Jiha)
- Diana Beinart (Representing Commissioner Preston Niblack)
- Louann Koziol (Representing Commissioner Thomas Foley)
- Hector German (Representing Speaker Adrienne Adams)
Agenda:
- Approval of the previous meeting minutes.
- Resolution recommending the Board of Directors approve the issuance of up to $1.6 billion in future tax-secured bonds.
Key Decisions:
- The minutes from the last meeting were reviewed and unanimously approved.
- The resolution regarding the bonds was presented, discussed, and unanimously approved by the committee members.
Action Items:
- None mentioned.
Adjournment:
- The meeting was adjourned.
This distilled format is significantly easier to read than the full transcript. Note that certain placeholders need filling based on the context and specifics of the meeting, but the overall structure provides a clearer summary.
In cases where the transcript is lengthy and exceeds character limits in ChatGPT, consider breaking it into smaller segments, summarizing each, and then compiling the results into cohesive meeting minutes.
Extracting Unstructured Information
ChatGPT’s capability to understand text allows you to extract unstructured information—data buried in paragraphs rather than neatly arranged in lists or tables. This is essential since unstructured data can be challenging to sift through.
Example for Extraction
Suppose we have this report excerpt:
The Port Authority of New York and New Jersey recently announced that air passenger numbers have surpassed pre-pandemic levels for the first time during the pandemic period. In November, the airports managed a total of 11 million passengers, representing 102 percent of November 2019’s pre-COVID volumes.
Using ChatGPT, you can ask:
- What was the latest passenger count mentioned?
- What date is referenced for this volume?
- How many passengers were there before?
- What was the date for the previous volume?
ChatGPT’s Extraction:
- The latest passenger count is 11 million passengers.
- The date referenced is November.
- The text does not specify how many passengers there were before the pandemic.
- The date for the previous volume is not mentioned.
The outcome presents a streamlined, easy-to-read format that can be utilized elsewhere. If the text lacks specific information, ChatGPT will indicate so.
How to Use ChatGPT for NLU
To effectively use ChatGPT’s NLU capabilities, it’s crucial to phrase prompts carefully to avoid generating irrelevant or “hallucinated” content. Hallucination occurs when the model fabricates information not grounded in the provided context. To minimize this, instruct ChatGPT to base its responses strictly on the information given.
Be cautious about specialized jargon or abbreviations that might be outside ChatGPT’s comprehension. If necessary, rephrase and regenerate your queries to improve the accuracy of the information extracted.
Summary
In this article, you have explored how to utilize ChatGPT as a powerful tool for natural language understanding. You have learned how to:
- Transform meeting transcripts into concise minutes
- Extract specific information from complex texts
By harnessing these capabilities, you can streamline various tasks, such as generating action item lists from emails and facilitating efficient documentation practices.