ABOVE THE CURVE: ADAPTING CHAT-BASED AI INTO PROFESSIONAL WORKFLOWS

Above the Curve

ABOVE THE CURVE: ADAPTING CHAT-BASED AI INTO PROFESSIONAL WORKFLOWS

Recently, CWEEL – SoCal chapter enjoyed its first Zoom presentation which was titled: “Above the Curve: Adapting Chat-Based AI Into Professional Workflows”. The presenters were Trevor Nelson and Kenza Nelson of Genie Prep. Genie Prep offers online FE (Fundamentals of Engineering) exam preparation and boasts over 1,000 students. Read more about Genie Prep at genieprep.com.

Key takeaways of the presentation are as follows:

After introductions were made, the presenter began by discussing Generative AI which was described as “…a type of artificial intelligence that focuses on creating or generating new content or data. This can be in the form of language, images, videos, and more.” Where Artificial Intelligence or AI is a more general category which can be any technology which performs tasks normally requiring human intelligence, Generative AI would be any one of several AI categories that can also include robotics, machine learning, or conversational AI.

The Generative AI commercial landscape is growing rapidly as more avenues are opening for AI-enhanced applications and solutions. However, one application has recently made news and has had a large impact on raising awareness of the current state of Generative AI and that is ChatGPT by OpenAI. ChatGPT is a conversational style application that allows a user to interact with a Large Language Model (LLM). ChatGPT is designed to understand and generate human-like text based on the input it receives.

Text Box: Figure 1 Poll - AI AwarenessAnyone interested in exploring ChatGPT can do so for free by heading over to openai.com and follow the links for setting up an account.

Diving deeper into the Large Language Models (LLM), by “adding one word at a time”, LLMs are trying to produce a “reasonable continuation” of the input text. The presentation describes “reasonable” as “what one might expect after reading billions of words of text from the Internet”.

A conversational application like ChatGPT is only one example of Generative AI in action. Other offerings that were discussed by the presenters include –

  • Alternative conversation applications such as Google’s Bard and Microsoft’s Bing Chat Bot application (which is powered by ChatGPT)
  • Magic ToDo by Goblin Tools (goblin.tools) – a “to do” list style application that is built on a LLM
  • Midjourney: a site using Generative AI to generate hyper-realistic art by simply providing a description of the art in a Chat Bot. (midjourney.com)

Following the introduction to Generative AI and more specifically, Chat Based AI, the presentation focused on the main topic of adapting the technology into your own workflows by first stressing that Generative AI should be looked at “as a tool and not a replacement”.

The slide titled, “The Need for Expert Guidance” stated that “AI models, by nature of their design, are prone to errors, which can range from simple data misinterpretations to substantial bias in decision-making. Human experts can identify and rectify these errors, ensuring the accuracy and fairness of the AI system’s output.” The meaning is clear, the technology is far from infallible. Output from the LLM’s should be evaluated and feedback provided to ensure that improvements are made, and more accurate results are produced over time.

AI at Genie Prep

The presentation illustrates the capabilities of Generative AI such as ChatGPT by highlighting some of the ways that Kenza and Trevor are applying the technology in their organization.

  1. Social Media – many of Genie Prep’s courses are available on YouTube and as they describe, “YouTube videos require a surprising amount of text”. Whether it is coming up with a title for the video or generating key works, ChatGPT has sped up their process by at least 3 X.
  2. Customer Support – ChatGPT facilitates conversations with students, mostly engineers who communicate with a different style than students of other disciplines. In addition, the application allows support to appropriately match tone when, for example, communicating by email, friendly verses academic.
  3. Source Material Search and Summarization – by matching source material such as industry guidelines as found in digital form (pdf’s) to the AI, Genie Prep can quickly locate and summarize information using natural language queries. For example, “What is the minimum footing thickness for a one-story, light frame construction according to the International Building
  4. Code (IBC)?” will return the requested information along with citations to the originating text.
  5. In addition, Genie Prep is currently planning an AI-based tutoring system which can help guide students to navigate the course material.

Several use cases were included which ranged from simple to advanced which could provide guidance to any professional looking for an entry point when considering using conversational AI such as ChatGPT. Examples included instructing the AI to write an email to a prospective client and using AI to help solve complex problems.

Summary

The presentation “Above the Curve: Adapting Chat-Based AI Into Professional Workflows” began with an introduction into the topic of Generative AI or a type of AI used to generate content or data such as text or images. It next went on to explore the current landscape of commercially available products that are all using AI right now which can either aid productivity, enhance learning or simply be used creatively. Today’s professional should be aware of the limitations and potential for error when using AI but the technology should be viewed as “Tool and not a Replacement” which has the potential to improve productivity.

The presenters illustrated the benefits of AI from the perspective of their experiences in incorporating AI into their own daily workflows at Genie Prep and providing guidance on how anyone can get started.

Deborah Smith, Liaison
CWEEL, SoCal
[email protected]