Large Language Model (LLM)
Class of generative models producing human‑like text.
How the laws define this
Regulators consistently define a large language model as a neural or AI model trained on vast text corpora or datasets to process, understand, and generate natural or human-like text. Despite this consensus on core technical characteristics, regulatory frameworks diverge significantly in their operational focus. The Data Protection Commission Guidance on AI and Large Language Models concentrates on mapping data protection obligations and preventing personal data disclosures during prompt-based inference. Conversely, the Government-wide Vision on Generative AI (Overheidsbrede visie generatieve AI) emphasizes public-sector risks, accessibility, and broad capabilities, while the National Artificial Intelligence Strategy 2024-2025 Action Plan (Ulusal Yapay Zeka Stratejisi) centers on establishing development milestones for Turkish LLMs and measurement frameworks. Further contrasting these regulatory approaches, the Fujian Province AI Industry Management Measures specifically regulates these models to promote local industrial integration. Compliance requirements thus depend heavily on whether an authority views these models through the lens of privacy risks, public-service adoption, national development goals, or regional industrial policy.
Synthesised from the 4 statutory definitions below. Each one is quoted in full, with its source.
How jurisdictions differ
What is distinctive in each definition, binding law in force first. Showing 3 of the 4 instruments that define this term; all are listed under Definitions below.
| Jurisdiction | Instrument | Status | What is distinctive |
|---|---|---|---|
| Fujian, China | Fujian Province AI Industry Management Measures | Superseded | Limits LLMs to those regulated under Fujian provincial measures for local AI industrial integration and SME subsidies. |
| Netherlands | Government-wide Vision on Generative AI | Adopted | Characterizes LLMs as an important subset of generative AI due to their broad capabilities, accessibility, public-sector relevance, and associated risks. |
| Turkey | National Artificial Intelligence Strategy 2024-2025 Action Plan | Adopted | Targets development milestones and measurement frameworks specifically for Turkish-language LLMs. |
Definitions (4)
A class of generative AI models trained on large text corpora to generate, complete or transform natural language; in the guidance LLMs are described in order to map data protection obligations to model design, training and runtime behaviours. The definition highlights LLMs' capacity for prompt‑based inference and risks such as unintended disclosure of personal data.
An important subset of generative AI characterized by large-scale neural models trained on vast text corpora to perform a wide range of language tasks. The vision highlights LLMs because of their broad capabilities, accessibility, and relevance to public‑sector use and risks.
A large language model (LLM) is defined as a neural network model trained on large corpora of text, typically used for language generation and understanding; the Plan specifically targets development milestones for Turkish LLMs and related measurement frameworks.
AI models trained on vast datasets capable of generating human-like text, specifically regulated under the provincial measures for local industrial integration.
Not to be confused with
AI conversation model refers to machine-learning models for generating or interpreting natural language, while LLM specifies large-scale models trained on vast text corpora.
Generative AI produces new content like text or images, whereas LLM is a specific subset focused on large-scale neural models for language tasks.
This term's definitions vary by instrument, while the related term provides a generic description of AI systems trained on vast text corpora to generate and analyze text.
Questions and answers
How does the Dutch vision describe LLMs in public sector context?
LLMs are an important subset of generative AI with large-scale neural models trained on vast text corpora for language tasks. The vision emphasizes their broad capabilities, accessibility, and relevance to public-sector use alongside risks.
What does Turkey's AI plan say about developing LLMs?
The plan defines LLMs as neural network models trained on large text corpora for language generation and understanding. It specifically targets milestones for Turkish LLMs and related measurement frameworks.
Related Terms
AI tools
Systems performing tasks that exhibit intelligent behaviour....
AI conversation model
Machine-learning models used to generate or interpret natural language....
Generative AI (GenAI)
AI that produces new content like text, images or code....
Large Language Models (LLMs)
AI systems trained on vast text corpora to generate and analyze text....
Runtime inputs
Inputs provided to a model during operation (prompts or user data)....