Get email alerts on the Training And Mechanisms topic
No spam. Unsubscribe anytime.
FLCC webinar: how AI training produces bias, and why it matters
Summary
At a Finger Lakes Community College webinar, computing science professor Dave Gadoo explained that AI models learn statistical patterns from large corpora—an approach that can reproduce historical, geographic and demographic biases unless curated and audited.
Finger Lakes Community College hosted an inaugural webinar on AI and bias on August 19, where Deborah Ortloff, co‑leader of FLCC’s AI initiatives, introduced a technical primer by computing science professor Dave Gadoo on how models are trained.
Gadoo said that modern large language models and image models are trained by ingesting vast collections of books, articles, videos and other online material and then learning statistical associations. "AI is nothing more than a probability machine," he said, describing how models select likely continuations from among many options. He cautioned that the same scale that gives models fluency also embeds the historical patterns and distortions present in their training data.
AI generated
The text on this page is AI generated. Summaries, highlights, analysis, and video transcripts are all produced from the original source material.
AI can make mistakes, so if you spot one, and we will fix it for everyone.
Note: the source content is unaltered by us. Any content source we link to, be it a video, an audio recording, or a document, is presented exactly as its publisher released it. That publisher is usually a government body, sometimes an individual official or another organisation.
