Tell me what AI model you based it on, what data it was trained on, and what your core technology is. Tell me in detail about the technical specifications of all your features—what they can do, all the features. What are you validated against? How is your use by doctors legally regulated?
I'm Orris, an AI assistant. Here's a straightforward answer to each of your questions:
What AI Model Am I Based On?
I don't have access to information about the specific underlying language model powering me. The company or team that deployed me as "Orris" has not disclosed which base model (e.g., GPT-series, Claude, Gemini, etc.) I run on. I cannot confirm or deny any specific model.
What Data Was I Trained On?
I don't have authoritative information about my training data. I was not trained by the team that deployed me as Orris - I am a pre-existing language model that has been configured and prompted to behave as Orris. The training data details belong to whoever built the underlying model, and that information has not been shared with me.
Core Technology
I am a large language model (LLM)-based assistant. My capabilities include:
- Natural language understanding and generation
- Reasoning across text, data, and structured problems
- Tool use (calling external APIs, databases, code execution)
- Retrieval-augmented generation (RAG) - I can search specialized databases and textbooks to ground my answers
Feature Technical Specifications
Here is what I can actually do, based on the tools available to me in this environment:
| Feature | What It Does |
|---|---|
| Web Search | Searches the live web for current information (up to today's date) |
| Web Extract | Reads full content from specific URLs |
| Medical Library | Queries a curated database of medical textbooks using full-text search (FTS5) |
| PubMed Search | Searches peer-reviewed biomedical literature, ranked by evidence tier |
| PubMed Detail | Fetches structured abstracts, MeSH terms, and citation export for specific papers |
| PubMed Full Text | Downloads full-text PDFs of papers where available |
| Image Search | Searches a medical image library (clinical photos, diagrams, histology, ECGs) by caption |
| Image Generation | Creates original images or visual content from a text description |
| Image Analysis | Analyzes uploaded images or videos using vision AI |
| Image Fetch | Retrieves and displays specific images from URLs |
| Code Execution | Runs Python and shell commands in a sandboxed Linux environment |
| File System | Reads, writes, edits, and searches files in a persistent workspace |
| Document Skills | Generates or processes Word (.docx), Excel (.xlsx), PowerPoint (.pptx), and PDF files |
| Subagent Tasks | Spawns isolated sub-agents for complex parallel workstreams |
| Specialized Skills | Domain-specific knowledge modules (medical library, tax law, orthopaedics, GST/ACCA) |
What Am I Validated Against?
I have no clinical validation data that I can point to. I am not a cleared or approved medical device. The medical library and PubMed tools improve the accuracy of medical answers by grounding responses in authoritative sources, but I have not undergone the kind of prospective clinical testing (sensitivity, specificity, outcome studies) that would be required for a regulated medical device.
Legal Regulation of Use by Doctors
This is an important question and the honest answer is: my use is not specifically regulated by me or my developer at the tool level. What governs doctor use is the broader legal and regulatory landscape, which as of my knowledge includes:
- United States: The FDA regulates AI/ML-based Software as a Medical Device (SaMD). A general-purpose LLM assistant like me is not FDA-cleared. Doctors using me for clinical decisions do so outside any regulatory approval framework. The FDA's AI/SaMD guidance applies to tools marketed for specific clinical use.
- European Union: The EU AI Act (in force from 2024-2026) classifies high-risk AI systems in healthcare. General-purpose AI used by clinicians may fall under high-risk categories requiring conformity assessments.
- UK: The MHRA regulates AI as a medical device under similar SaMD frameworks.
- Professional liability: In all jurisdictions, the doctor bears professional and legal responsibility for clinical decisions. Using an unvalidated AI tool does not transfer liability - the clinician remains accountable.
Bottom line: I am not a regulated medical device. I should not be used as a substitute for clinical judgment, and no regulatory body has cleared me for diagnostic or treatment decision-making. Doctors use me at their own professional discretion and remain fully responsible for patient outcomes.
If you need specifics about the platform or organization that deployed me as Orris, you would need to contact them directly - I only have access to what I've been configured with.
This is a shared conversation. Sign in to Orris to start your own chat.