large language model (LLM)
What It Means
Large Language Models are AI systems trained on massive amounts of text data that can understand and generate human-like language. They come in two main types: generative models that create new text (like writing emails or answering questions) and discriminative models that classify or analyze existing text (like detecting spam or determining if content was AI-generated).
Why Chief AI Officers Care
LLMs represent both significant opportunity and risk for enterprises, as they can automate content creation and analysis at scale but also introduce concerns around data privacy, accuracy, and regulatory compliance. The distinction between generative and discriminative LLMs matters for governance because they have different risk profiles - generative models can create misleading content while discriminative models are used for content moderation and detection.
Real-World Example
A financial services company uses a generative LLM to draft personalized customer communications and investment summaries, while simultaneously deploying a discriminative LLM to scan all outgoing content to ensure it meets regulatory requirements and wasn't generated inappropriately by unauthorized AI tools.
Common Confusion
People often think all LLMs work the same way, but generative models (like ChatGPT) that create new content have very different applications and risks than discriminative models that analyze and classify existing text.
Industry-Specific Applications
See how this term applies to healthcare, finance, manufacturing, government, tech, and insurance.
Healthcare: In healthcare, LLMs are increasingly used for clinical documentation, patient communication, and diagnostic support, but...
Finance: In finance, LLMs are being deployed for investment research summarization, regulatory document analysis, and automated c...
Premium content locked
Includes:
- 6 industry-specific applications
- Relevant regulations by sector
- Real compliance scenarios
- Implementation guidance
Technical Definitions
NISTNational Institute of Standards and Technology
"a class of language models that use deep-learning algorithms and are trained on extremely large textual datasets that can be multiple terabytes in size. LLMs can be classed into two types: generative or discriminatory. Generative LLMs are models that output text, such as the answer to a question or even writing an essay on a specific topic. They are typically unsupervised or semi-supervised learning models that predict what the response is for a given task. Discriminatory LLMs are supervised learning models that usually focus on classifying text, such as determining whether a text was made by a human or AI."Source: AI_Assurance_2022
Related Terms
Discuss This Term with Your AI Assistant
Ask how "large language model (LLM)" applies to your specific use case and regulatory context.
Start Free Trial