Leading Insights Blog

Revolutionizing Conversations: ChatGPT’s Potential Uses and Limitations in Conversational AI

By Daniella Dorio

Introduction

As technology continually pushes the boundaries of human ingenuity, conversational artificial intelligence (AI) has emerged as a fascinating and transformative force. It is increasingly evident that AI can be leveraged to assist users’ lives in everyday tasks. These natural language processing (NLP) advancements, like ChatGPT, have revolutionized the way we interact with virtual assistants. This article explores Generative Pre-Trained Transformer (GPT) 4 architecture, prompt engineering, potential uses, and limitations.

GPT 4 Architecture

Conversational AI has become a highly sought-after component of human-computer interaction, with language models playing a crucial role in this development. NLP has undergone significant advancements with the emergence of large language models (LLM). NLP is an AI subfield “concerned with giving computers the ability to understand text and spoken words in much the same way human beings can.”(1) These advancements gave way to ChatGPT’s model version GPT 4 architecture, employed on an unprecedented and disruptive scale of 175 billion parameters.(2) Given the speed and scale of AI technological advancements, there is evidently a monumental shift in human-computer interaction.

Developed on November 30, 2022, by OpenAI, ChatGPT is a groundbreaking advanced language model capable of generating responses remarkably similar to those of human conversations.(3) The “GPT” in ChatGPT stands for Generative Pre-trained Transformer, a type of neural network that learns context and can perform data analytics, typically on language data.(4) GPTs are “generative” because they can create new data and are “pre-trained” because they leverage existing models instead of creating new ones. The conversational nature of using ChatGPT highlighted the need to effectively craft optimal prompts that will elicit desired responses.

Prompt Engineering

To optimally use conversational AI, such as ChatGPT, it is important to understand prompt engineering. Prompt engineering is the process of describing a task that a user wants an AI chatbot to accomplish.(5) By understanding how AI processes the query, a prompt engineer can manipulate responses to achieve the intended results. For example, if an AI bot is asked “what is the weather,” the results can widely vary from indoor and outdoor temperatures, location, and any additional desired information, such as humidity or air quality. By refining the prompt to “what is the outdoor temperature, humidity, and air quality in Washington, DC” the chatbot can more accurately respond with the requested information. Crafting prompts encompasses technical skills and contextual awareness to construct optimal queries and responses. Once achieved, ChatGPT can be harnessed across a range of functions, including simplifying complex ideas, code composition, and résumé evaluation.

ChatGPT Potential Uses

  1. Simplify Complex Ideas
    Subject Matter Experts (SME) must simplify their technically robust findings into digestible, easy-to-understand topics. For example, if a client is working with AI and does not understand what LLMs are, they can ask ChatGPT to “explain what a large language model is, at a high-level.” Or they can ask ChatGPT to “rewrite this sentence to remove unnecessary jargon” to provide concise explanations. In doing so, users can simplify complex ideas and create persuasive presentations.
  2. Code Composition Development
    ChatGPT can additionally be applied as a code composition resource. Users can instruct ChatGPT to write code in specific languages by inputting, “write code to concatenate Employee and Salary data tables in Python,” for example. Users must remember that ChatGPT does not understand context or how the data model works, which should be considered because new code must properly integrate with existing datasets. ChatGPT should be used as a supplementary resource similar to developers leveraging Stack Overflow to ask coding questions and investigate syntax best practices. Nonetheless, ChatGPT serves as a complimentary resource for code composition.
  3. Résumé Review and Evaluation
    ChatGPT can write, edit, and read résumés, making the AI chatbot an invaluable resource for recruiting staff. ChatGPT can assist with content analysis, keyword matching, and overall impressions of the applicant’s résumé. For example, ChatGPT can assess the job description’s most common words and phrases and prioritize résumés based on their keyword frequency. To optimize ChatGPT for résumé evaluation, recruiters and users should consider uploading the job description, sample résumés, and other materials to maximize context and relevance. In turn, recruiters can increase efficiency, objectiveness, and consistency in the résumé evaluation process.

Limitations

With every new technology, there are always areas of potential improvement. While ChatGPT provides powerful conversational capabilities, users should also be mindful of its limitations and exercise caution to relying solely on its responses, especially in sensitive domains. Limitations include virtual assistants’ lack of contextual understanding, sensitivity to prompt engineering, and ethical and legal considerations. These considerations include a lawyer using ChatGPT to write a legal brief, in which the chatbot included fictional legal case references.(6) There are additional security concerns, including bad actors using ChatGPT to create phishing e-mails.(7) Users should always verify information using reliable sources and consider the responsible use of AI technologies.

Conclusion

The emergence of conversational AI has heralded a new era in human-computer interaction, transcending the boundaries of what was once thought possible. As technology continues to advance, AI’s capacity to comprehend and engage in meaningful conversations with users has become increasingly sophisticated. In fact, due to the influx of AI and virtual assistants in recent years, as well as the nuance required to create contextually relevant queries, many companies have created prompt engineering jobs to train an “emerging crop of AI tools to deliver more accurate and relevant responses to the questions real people are likely to pose,” thus creating new job markets as well as disrupting existing ones.(8) Language models’ wide-ranging applications make it a powerful tool for democratizing AI technology for average consumers. However, addressing ethical and legal considerations is critical to effectively harness AI’s full potential while ensuring its responsible integration into our lives.

Sources

(1) What is Natural Language Processing? | IBM
(2) Introducing ChatGPT (openai.com)
(3) ChatGPT — Release Notes | OpenAI Help Center
(4) The A to Z of Artificial Intelligence | Time
(5) What is prompt engineering? Definition + skills | Zapier
(6) ChatGPT: US lawyer admits using AI for case research – BBC News
(7) Is ChatGPT a cybersecurity threat? | TechCrunch
(8) How to Get a Six-Figure Job as an AI Prompt Engineer | Time
(9) How to write an effective GPT-3 or GPT-4 prompt | Zapier

To top