Epic AI Fails

Free Mini Conference and L&D Happy Hour 
in Chicago

When:

Thursday, October 1st

3:30 - 7:30 p.m. CDT

Where:

Clark Street Ale House

742 N Clark St, Chicago IL, 60654

What:

Before you invest more time in AI, learn from true stories about what doesn't work.

Join us October 1 as L&D professionals discuss AI approaches that went sideways, and the practical lessons you can take from their sometimes-painful experiences. Bring your own stories to share and connect with peers during a free happy hour!

No selling and no cost! Just learning and connection (plus drinks and snacks) with L&D peers.

These engaging speakers will each share a 20- to 30-minute session:

Chad Udell of SparkLearn

Technically Correct and Completely Useless: What Frontline AI Failures Actually Teach Us

 Emily Hoffmann of Fredrickson Learning

The Illusion of Progress: What AI at Work Gets Backward

Shylee Garrett of Meta

Before You Feed the AI: What L&D Needs to Know About Knowledge Quality

Read on below to learn more about the sessions.

The sessions:

Technically Correct and Completely Useless: What Frontline AI Features Actually Teach Us

Chad Udell - CEO, SparkLearn



Most AI failures in frontline learning don't look like failures. Nothing blows up and the output reads just fine. The SOP summary is "accurate." The generated micro-course is well structured but just a little "off". The chatbot answers the question confidently, but just wrong enough to erode trust.

I've spent the last couple of years watching organizations bolt AI onto messy content, shared tablets, spotty connectivity, and governance nobody wrote down, then act surprised when it goes sideways.  

AI is a layer, not a foundation, and it will happily remix your chaos at scale. You'll leave with a short checklist of the questions to ask before your next AI pilot, so your fail story stays a lot less epic than mine.

The Illusion of Progress: What AI at Work Gets Backward

Emily Hoffmann - Fredrickson Learning



Human factors research in aviation and medicine has explored what happened when automation took over parts of expert work: pilots losing manual skills they still needed, surgical residents who stopped getting hands-on practice, clinicians following a wrong recommendation because it came from the system. Much of it maps onto questions L&D is asking now. How does someone become a senior designer if AI does the junior work? What is left of the job? How would we know if AI is making our decisions worse? This talk goes through the epic fails those fields already had, what they learned from them, and whether the comparison to our situation holds.

Before You Feed AI: What L&D Needs to Know About Knowledge Quality

Shylee Garrett - Meta



More content doesn’t necessarily mean better AI. As organizations rush to integrate generative AI into learning and knowledge management, many are overlooking a critical question: Is the knowledge we’re feeding AI actually ready to be used?

From outdated training materials and duplicated procedures to conflicting sources of truth, existing content management practices can undermine AI’s ability to deliver accurate, reliable answers.

This session explores why traditional approaches to learning content and knowledge management need to evolve in an AI-driven workplace. We’ll examine common mistakes organizations make when preparing content for AI, the importance of content quality and governance, and how L&D teams can help build a more trustworthy knowledge ecosystem.