Summary
The webcast delves into ChatGPT's capabilities, discussing the free and subscription versions along with generative AI usage for work tasks. It explores foundational models in deep learning, large-scale model advancements, and AI applications in public feedback synthesis. The session addresses bias in AI models, emphasizing accountability for potential misinformation and its implications for public discourse. Custom GPT models for specific tasks like zoning ordinance reviews are showcased, with a demonstration highlighting the model's ability to interpret complex documents tailored for its use. The importance of quality inputs, model constraints, and contextual understanding when working with AI models is stressed throughout the discussion.
Chapters
Introduction
Polling Questions and Poll Results
Technology Review Considerations
Ethical Considerations
Strategies in Practice
Introduction to ChatGPT4 Homepage
Custom GPT Model for Zoning Ordinance
Understanding Variance Application
Running the GPT Model
Different Variance Request Scenario
Exploring GPT Limitations
Introduction
Introduction to the webcast, moderator, technical help instructions, guest host, registration information, social media channels, and Q&A guidelines.
Polling Questions and Poll Results
Discussion and polling questions regarding the free version of chat GPT, subscription version of AI software, and generative AI usage for tasks at work.
Technology Review Considerations
Overview of foundational models, deep learning revolution, large-scale models, data sources, application of AI in public feedback synthesis, image and video generation advancements, and open-source ecosystem development.
Ethical Considerations
Exploration of bias in AI models, implications for planning practice, impact on public discourse, potential misinformation, and accountability for AI-generated content.
Strategies in Practice
Discussion on generating custom GPT models, constraints, fine-tuning techniques, staff report creation demo using GPT, and resources for planners.
Introduction to ChatGPT4 Homepage
The speaker introduces the ChatGPT4 homepage where users can input queries similar to a Google search and receive results. The custom GPT model discussed is designed for reviewing zoning ordinances in North Carolina for a variance application.
Custom GPT Model for Zoning Ordinance
The speaker explains the customized PDF application loaded into the GPT, tailored for a variance application in a specific format suitable for the GPT to interpret. The model is designed to simplify the reading process for the GPT.
Understanding Variance Application
Details about a variance application in the R2 Zone for a single-family zoning district are provided, where an individual seeks variances for building height and setback regulations to construct a large residence. The extreme nature of the variance request is highlighted.
Running the GPT Model
The GPT model is executed to assess the variance request, and the generated report recommends denying the request due to substantial deviations from the zoning ordinance and failure to meet criteria.
Different Variance Request Scenario
A new variance request scenario is presented, focusing on a front setback variance due to property topography. The GPT analysis recommends approving the request based on genuine hardship caused by topographical challenges.
Exploring GPT Limitations
The speaker discusses the limitations of the GPT model, highlighting the importance of quality inputs, narrowing the model focus, and emphasizing the significance of context in working with AI models.
Get your own AI Agent Today
Thousands of businesses worldwide are using Chaindesk Generative
AI platform.
Don't get left behind - start building your
own custom AI chatbot now!