EMA & HTA Guidance on Use of Large Language Models
- Date
- September 26, 2024
- Category
EU Regulations
- Description
On 29 August 2024 the EMA and HTA jointly issued guidance titled “Guiding principles on the use of large language models in regulatory science and for medicines regulatory activities”. See www.ema.europa.eu/en/documents/other/guiding-principles-use-large-language-models-regulatory-science-medicines-regulatory-activities_en.pdf
A statement by the EMA and HMA, on 5 September 2024, stated that large language models (LLMs) have “enormous transformative potential” for medicines regulation, and can be used for a variety of tasks, from generating and translating text to summarising documents.
The guideline introduction covers:
- What are large language models?
- Use of large language models
- Safe and responsible use of LLMs
The guidance continues with:
General ethical considerations
User principles
Organisational principles
The guideline concludes that “In this fast-evolving field, it has never been more important to provide a structured approach to integrating advanced AI tools into regulatory processes to safeguard public health and maintain trust in medicines regulatory authorities.”
The EMA and HTA also released a one-page infographic factsheet on the “4 principles for safe and responsible use of LLMs” on 1 September 2024. See https://www.ema.europa.eu/en/documents/other/factsheet-four-principles-safe-responsible-use-large-language-models_en.pdf