How we use Artificial Intelligence

We believe AI (Artificial Intelligence) and the use of large-language models (LLMs) are important technologies that should not be understated. However, complacent use of AI undermines our third principle, "Transparency over complexity", and can lead to models and estimations that are difficult to interpret. healthmodel.org uses AI as a supplement to our modelling protocol, incorporating it as a tool to build models and increase transparency.

We use AI in the following ways:

1. Allowing LLM agents to build or adapt models by accessing our model building tools via an Model Context Protocol (MCP)
2. Allowing LLMs to interface with users and models as chat-bots, explaining complex models to the right audience based on their targeted questions
3. Allowing LLMs to populate sparse data with plausible assumptions so that users can begin understanding the models
4. Using LLMs to assist with programming the infrastructure that underlies our platform, albeit with heavy guardrails and testing