AI Discussion with ChatGPT – Introduction

A forthcoming 7-part series of articles about Artificial Intelligence (AI) began as a personal educational exercise with three primary objectives.

First, to expand this writer’s general knowledge of AI beyond: 1) AI-related articles in news publications like the New York Times and Wall Street Journal: 2) more in-depth AI coverage and analysis in newsletters such as Zvi Mowshowitz’s Don’t Worry About the Vase (paid subscription optional); and 3) insurance industry specific AI coverage such as Faegre Drinker’s AI newsletter.

This AI knowledge study builds upon this writer’s prior interest in, study of, and writing about knowledge management and the emerging social network technologies during the mid-2000s.

Second, a dialogue with ChatGPT Version 4.0 specifically about AI offered a focused opportunity to experiment with a fundamental AI skill set (“prompt engineering”) beyond two prior attempts within this writer’s own subject matter expertise:

  1. “A Structured Settlement Interview with ChatGPT” 
  2. “ChatGPT Discusses Qualified Settlement Funds (QSFs)” 

As highlighted in a recent Wall Street Journal article (subscription required): “Prompt engineering—crafting the sharpest queries to get the best answers out of ChatGPT or another generative AI chatbot—is a new type of job” – with salaries of up to $200,000 per year.

Third, selective, expanded utilization of ChatGPT (currently the most popular generative AI chatbot) provides this writer a better understanding and appreciation of ChatGPT’s current shortcomings – of which there are several.

One of the most important limitations of ChatGPT Version 4.0 is time specific. Its last general update (as of December 2023) occurred in April 2023. In the warp-speed world of AI, that time-lapse seems like eons and must be considered – both in formulating prompts and when reading ChatGPT responses.

Another problem with ChatGPT, and presumably all chatbots, is inaccurate and untrustworthy data – what AI researchers call generative AI “hallucinations.” This problem is one reason many companies do not allow workers to use consumer versions of chatbots for work-related activities.

To address the data accuracy problem of chatbots, a recent New York Times article (subscription required) reported: “a consortium of companies has developed standards for describing the origin, history and legal rights to data. The standards are essentially a labeling system for where, when and how data was collected and generated, as well as its intended use and restrictions.”

Another problem with chatbots is privacy – or potential lack thereof. Be certain not to share personal or business or otherwise private information when using chatbots. Privacy concerns represents another reason why many companies do not allow workers to use consumer versions of chatbots for work-related activities.

How Should You Use Chatbots?

What are AI chatbots good for now? Another New York Times article titled How Should I Use A.I. Chatbots like ChatGPT” identifies six tasks large language models are already good at:

  1. As a personal tutor to explaining concepts at multiple difficulty levels.
  2. Editing and constructing criticism for articles and papers.
  3. Overcoming writers’ block or preparing for a challenging project.
  4. Rehearsing for difficult off-line assignments.
  5. Summarizing large amounts of text.
  6. Coding, even if you don’t know any programming languages.

Which chatbot should you use? As of December 2023, OpenAI ChatGPT still appears to be the leading chatbot. However, Google recently announced a series of upcoming upgrades to its chatbot Bard. Microsoft launched its chatbot Copilot as Bing Chat on February 7, 2023. Other chatbots and chatbot upgrades will continue to enter the market in the weeks and months ahead.

AI SERIES OUTLINE AND OBJECTIVES

Having completed a personal AI tutorial, Independent Life, this writer’s employer, has agreed to publish the resulting articles as a forthcoming series featuring this writer’s prompts and ChatGPT Version 4.0’s unedited responses as follows:

  1. Part 1 – AI Vocabulary and Definitions
  2. Part 2 – History of AI
  3. Part 3 – The Current AI Market
  4. Part 4 – OpenAI and ChatGPT
  5. Part 5 – AI Tools
  6. Part 6 – AI Legal and Regulatory Developments
  7. Part 7 – AI’s Future and AI Issues

The primary objectives for publication: 1) to provide structured settlement and settlement planning professionals with an introduction to AI; 2) to encourage them to learn more about AI; and 3) to suggest they (continue to) experiment with personal use of generative AI chatbots including ChatGPT, Bing Chat Enterprise, Google Bard, Microsoft Copilot among others.

Note: neither Independent Life nor this writer is responsible for any false or inaccurate information generated by ChatGPT responses within any of the above-listed articles.

Share