As part of yesterday’s webinar, we asked AI to identify ten AI-related terms that are frequently misunderstood. Here are the terms it gave us.

The webinar was titled AI in the Courtroom: How Generative AI is Changing the Way Legal Teams Go to Court (sponsored by Exterro and available here for on-demand viewing) and we had a fantastic panel that I moderated: Martin Tully, Partner at Redgrave LLP; Kelly Twigger, Principal of ESI Attorneys and Bryant Bell, Lead Product Marketer – eDiscovery & AI at Exterro.

During our prep call, we discussed the idea of identifying common AI-related terms to highlight some of those misunderstandings to the audience and using AI to do so. So, I put this prompt into ChatGPT: “Please identify 8 to 10 AI-related terms that are most often misunderstood by those using AI”.

Being the overachiever it is, ChatGPT gave me ten AI-related terms that are frequently misunderstood and five “honorable mentions”. Here are the ten AI-related terms that are frequently misunderstood and why, according to ChatGPT:

1) Artificial Intelligence (AI)

Many people use AI as a catch-all term for any advanced software. In reality, AI encompasses a broad range of technologies designed to perform tasks that typically require human intelligence, including machine learning, computer vision, natural language processing, and expert systems.

2) Generative AI

Often confused with AI in general, generative AI refers specifically to systems that create new content such as text, images, audio, video, or code. Not all AI systems are generative, and not all AI applications involve content creation.

3) Large Language Model (LLM)

Many users assume an LLM “understands” information like a human. An LLM is a statistical model trained on vast amounts of text to predict the most likely next word or token. Its outputs can appear intelligent without reflecting true understanding or reasoning.

4) Hallucination

A hallucination is not simply an AI mistake. It occurs when an AI system confidently generates information that is inaccurate, fabricated, or unsupported by facts. Hallucinations are a byproduct of probabilistic language generation rather than a software bug that can be completely eliminated.

5) Training Data

Users often believe AI models directly access or “remember” all information from their training data. In reality, training data is used to develop patterns and relationships within the model. The model does not store or retrieve documents in the way a database does.

6) Reasoning

Recent AI models are frequently described as “reasoning models,” leading some users to believe they think like humans. AI reasoning generally refers to the model’s ability to perform complex, multi-step analysis and problem-solving, but it is still fundamentally based on mathematical pattern recognition rather than human cognition.

7) Agentic AI

One of the most misunderstood emerging terms. Agentic AI refers to systems that can plan, make decisions, use tools, execute tasks, and pursue objectives with varying levels of autonomy. It does not necessarily mean fully autonomous or self-governing AI.

8) Context Window

Users often assume AI remembers everything discussed in a conversation. A context window is the amount of information a model can consider at one time. Information outside that window may be forgotten or inaccessible unless specifically retained through other mechanisms.

9) Retrieval-Augmented Generation (RAG)

Frequently mistaken for model training. RAG is a technique that allows an AI system to retrieve information from external documents or databases at the time of a query and use that information when generating a response. It helps improve accuracy without retraining the model.

10) Fine-Tuning

Many organizations believe fine-tuning means teaching a model new facts. While it can improve a model’s performance on specific tasks or domains, fine-tuning primarily adjusts model behavior and outputs rather than serving as a simple knowledge update mechanism.

I can see all of them as being regularly misunderstood; in fact, I recently wrote a post on why hallucinations are misunderstood and why they won’t go away in LLMs.

AI got me this far, so I decided to take a step further and load these ten terms and load the output of the ten terms into NotebookLM and have it create a slide graphic that I could include into a slide deck. Here is what NotebookLM came up with:

Here is that same image, full-sized.

So, what do you think? How many of the ten AI-related terms that are frequently misunderstood have you misunderstood? Please share any comments you might have or if you’d like to know more about a particular topic.

Note: This post contains AI-generated content. Duh! 😉

Image created using DALL-E 3, using the term “robot lawyer wearing a suit and looking at a workstation confused”.

Disclosure: Exterro is an Educational Partner and sponsor of eDiscovery Today

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.