Hello newsletter subscribers! It has been awhile. This newsletter will be AI focused this time around. 

Generative AI in Statistical Analysis

I just published a new blog post on responsible generative AI usage in statistical analysis. It is focused on chatbots and how to use them effectively to meet your analytical goals. While these tools can provide assistance with coding and statistical approaches, they can give incorrect or incomplete information (as we all know at this point). The recommendations are: (1) develop and understand your statistical models before coding/implementation; (2) know clearly what you want and write prompts that reflect that; (3) review all output (explanations and code) and make sure you understand it; and (4) verify chatbot output with reliable sources. Starting with simple tasks and incrementally increasing complexity is a good general best practice to follow. Usage of generative AI is becoming more common in statistical analysis, so let’s make sure it helps us and doesn’t lead us to incorrect methodologies or solutions.

Mindrouter and Other UI Resources

If you are not already using University of Idaho’s Mindrouter for generative AI tools, this is worth checking out. This is an internal tool for UI affiliates to access over 60 open weight models. Any questions asked, information provided or data uploaded remain internal to the University of Idaho. They also have a regular blog you can subscribe to or read the archives. I found this post on setting up a local agent interfacing with Mindrouter to be particularly helpful.  

Mindrouter is one of the most secure ways to interact with generative AI tools without compromising data ownership. When working with private generative AI tools like Claude, the terms of service for those tools and companies govern how your data may be used by the company that owns the tool(s). See the University of Idaho’s IT policy for identifying and protecting high risk data.  

If you are outside of the University of Idaho, this tool is not directly available to you. However, Mindrouter infrastructure is available for IT administrators anywhere to consider implementing at their own institutions. In their words:

MindRouter is a production-ready LLM inference load balancer and translation layer that fronts a heterogeneous cluster of Ollama and vLLM inference backends. It provides a unified OpenAI-compatible API surface with native Ollama compatibility, fair-share scheduling, per-user quotas, full audit logging, and real-time GPU telemetry.

—Mindrouter website

The University of Idaho library has also assembled many useful resources for generative AI including some AI 101 information, guidance on responsible usage of this technology, resources for instructors, and links to UI policy. IT has also put together a short guide for choosing the right AI tool to meet your needs. 

Upcoming Events

Data Carpentries

Git, R, and Unix for Reproducible Scientific Analysis

• A six-session Software Carpentries workshop covering Unix, shell scripting, Git, and R 

• November 3–19, 2026 

• Tuesdays & Thursdays 3:00–5:30 PT 

• IRIC 352 or Zoom

Software Carpentry aims to teach researchers basic concepts, skills, and tools for working with data to get more done in less time, and with less pain. This hands-on workshop introduces the Unix shell, Git version control, and R — the core toolkit for reproducible scientific computing. Topics include navigating the command line, writing shell scripts and loops to automate repetitive tasks, tracking project history with Git and GitHub, and wrangling and visualizing data in R using tidyverse and ggplot2. A key feature of this workshop is integrating AI tools (Claude, ChatGPT) as coding partners: participants learn to write effective prompts, critically evaluate AI-generated code, and iterate toward working solutions.

The workshop is aimed at graduate students and researchers but is open to all; no prior programming experience is required. Participants must bring a laptop with a Mac, Linux, or Windows operating system on which they have administrative privileges. For more information, visit the IMCI website.

AI symposium

• Thursday, October 29, 2026

• 9:00 am to 4:30 pm

• International Ballroom, Bruce Pitman Center and Zoom

UI’s Center for Excellence in Teaching and Learning (CETL) is organizing a one-day AI Symposium on October 29 (9:00 am – 4:30 pm), titled "Navigating AI Together: Opportunities, Challenges, and Responsibilities." This is an in-person event with an online option for individuals not on the Moscow campus. The goals of the symposium are to (1) support informed and critical engagement with AI; (2) Build practical awareness through shared examples and use cases; (3) examine implications for teaching, learning, research, and policy; and (4) foster dialogue and shared learning across the campus community.

More information can be found on the symposium SharePoint site, including the link for free registration.