Large Language Model (LLM)
An AI system trained on vast amounts of text to predict and generate human-like language. Best understood as a concept-transformation engine rather than a knowledge database.
Definition: An AI system trained on vast amounts of text to predict and generate human-like language. Best understood as a concept-transformation engine rather than a knowledge database.
A Large Language Model (LLM) is an AI system that has been trained on enormous amounts of text data to learn statistical relationships between words and concepts. Models like GPT, Claude, and Gemini are examples of LLMs.
How They Work
At their core, LLMs are advanced prediction machines. When you provide a prompt, the model does not "think" or "know" the answer—it calculates the most probable next token (word or word-part) based on patterns learned during training.
The "T" in GPT stands for Transformer, which describes the model's core function: transforming concepts from one form to another.
What This Means for Research
Understanding this architecture is key to using LLMs effectively:
- They are transformation engines: Excellent at restructuring, rephrasing, and reformatting information you provide
- They are not knowledge databases: They generate plausible text, not verified facts
- They excel at pattern matching: Strong at identifying common issues because they have seen similar patterns in training data
- They struggle with novelty: Less effective at discovering truly new insights that require contextual understanding
Practical Implication
Treat an LLM as a powerful autocomplete and concept connector, not as a perfectly factual source. The best results come from giving it structured tasks to perform on information you provide, rather than asking it to generate knowledge from scratch.
Related Terms
Hallucination
When an AI model generates plausible-sounding but factually incorrect or fabricated information. A natural consequence of how LLMs predict probable text rather than verify truth.
Retrieval-Augmented Generation (RAG)
A technique that enhances LLM responses by first retrieving relevant information from a specific knowledge base, then using that information to ground the model's output.
Mentions in the Knowledge Hub
This term is referenced in the following articles:
Building a Research Career in the Age of AI
AI is transforming what researchers do daily, but it amplifies rather than replaces the core value researchers provide. Understanding which skills remain essential and how to grow them is critical for career development in this changing landscape.
Evaluating AI Research Tools: A Durable Framework
The AI landscape changes weekly. Rather than chasing specific tools, you need a durable framework for evaluating any platform against principles that will not change: privacy, transparency, portability, and reproducibility.
What AI Can and Cannot Do for UX Research
AI is not going to take your job, but it is absolutely going to change it. Understanding what LLMs actually are, and are not, is the foundation for using them effectively.
Advanced AI Techniques for Research
Beyond basic prompting, there are techniques that dramatically improve AI reliability: structured communication, using notes over transcripts, treating models as a committee of raters, and understanding when RAG or fine-tuning makes sense.
Research Tools and the ResTech Landscape
The research technology (ResTech) landscape has exploded with specialized tools for every phase of the research process. Understanding this ecosystem helps you choose tools that amplify your capabilities without creating dependency or replacing critical thinking.
AI-Assisted Thematic Analysis: A Practical Workflow
The biggest mistake teams make with AI is treating it like a magic black box. Here is a complete, reliable workflow for using LLMs as research assistants while maintaining critical human oversight.
Synthetic Research Data: Automated Walkthroughs vs. Fake Users
AI agents can simulate a logical user journey, but they cannot simulate the messiness of human behavior. Where to draw the line between useful stress-testing and dangerous fabrication.
Ethics and Data Privacy in UX Research
Our work gives us the privilege of entering other people's lives. With that privilege comes profound ethical responsibility, especially in the age of AI tools and cloud-based analysis.
Qualitative Thematic Analysis: From Codes to Insights
Transform interview transcripts and observation notes into actionable themes through systematic coding. The difference between an opinion and a finding is whether two people agree.
Cross-Cultural Research: Internationalization & Localization
Translating a UI is easy; translating an experience is hard. How to use back-translation and local partners to avoid cultural blind spots.