AI Explained – Part 1

This series of posts aims to demystify Artificial Intelligence (AI) and ChatGPT, a technology that has fascinated the world but also confused many with its unfamiliar terms and acronyms. We will introduce you to the basics and applications of this technology from a business perspective.

AI has a long history, dating back to the 1960s when the mathematical foundations were laid and even earlier to Aristotle’s time when logic as a process was created. In the 1990s, a huge leap forward occurred when different technologies such as machine learning, neural networks, natural language processing and others were combined to create new possibilities. Natural Language Processing (NLP), which is essential for ChatGPT, advanced significantly in 2017 with Google’s Transformer deep learning models. That’s what the T in GPT means: Generative Pre-Trained Transformer.

One of the reasons why ChatGPT is so popular is its natural way of interaction. We text or chat online with others frequently. By using this interaction, the user already knew how to use ChatGPT. In fact, the chatbot experience was designed to slow down the appearance of text in the chat window to make it seem like someone was actually typing on the other side. Otherwise, it would look like most other online messages, appearing all at once.

Most people use ChatGPT to solve one-step problems. We will explore how users are using prompt engineering to design and execute multi-step processes. We will also look at AI orchestration where you can use different types of AI functions from one workflow. AI orchestration will be a major growth area beyond ChatGPT. We already see companies integrating AI into Zapier and Power Automate workflows, where AI is used to detect, evaluate, and achieve results. There is still a human in the loop but maybe only one instead of 15.

We will also examine the role of “human in the loop” design in your AI processes. Some companies may be tempted to use AI as a way to reduce staff but that may be a hasty decision. We will explain how “human in the loop” design can mitigate risk in situations where a wrong outcome could have severe consequences. For example, what if an AI misdiagnosed a patient or an AI bot for health insurance gave incorrect information about a policy to a patient? This would ruin any trust the customer would have in your company.

We will also cover the distinct types of AI functions in upcoming segments. We will discuss Generative AI, which ChatGPT belongs to, and how it can help your company create new content. We will also explore the knowledge management cycle and how AI can enhance your organization’s learning and adaptability. We will also introduce other types of AI such as Content AI and how they can enable new automation possibilities in your organization. Lastly, we’ll look at AI adjacent technologies like Microsoft Fabric and how these technologies can help support your AI journey.

Stay tuned to learn more about this fascinating field of technology.