Presenter offers practical mitigation: inclusive models, source control and user auditing

Sep 21, 2026

The webinar closed with practical guidance: use inclusive LLMs, sanitize and audit sources, run controlled local models for sensitive data, and apply prompting and thumbs‑up/down judiciously; attendees were told to avoid putting PII into AI inputs.

The full story

Gadeau offered concrete steps attendees can use immediately: prefer inclusive models where available (he cited a startup called Latimer as an example), audit or sanitize training data, and run local LLM instances when privacy is a priority. He recommended specific user actions: check outputs, correct prompts when biased results appear, and use review workflows rather than accepting AI results uncritically.

Free newsletter

Sign up to keep reading.

Your free newsletter brings you what your city council, county, and school board discussed and decided.

or
Already have an account? Sign in