Speaking

AI didn’t break teaching. It showed us what a finished file was never proving.

“The future is already here — it’s just not evenly distributed.” — William Gibson

I’ve spent four years rebuilding my finance courses for AI, little by little at first and then from the ground up. Today AI is allowed on every assignment, and every arc still ends with the AI off. I’m not selling a finished answer. I share the whole journey, failures included, so your faculty can skip the mistakes I’ve already made and start redesigning on Monday.

Chris Wasden

About Chris

Chris Wasden, EdD, is Professional in Residence, Finance, at the Woodbury School of Business, Utah Valley University, and founder of Crucibl Methodology, LLC. He teaches finance with no textbook and AI on every assignment, and he designs backward from what students must do on their own. He helps educators and leaders rethink, restructure and reauthor their courses and assessments for the AI era.

EdD, George Washington University (2010) · MBA, UCLA Anderson School of Management (1987) · BS Accounting and BA Asian Studies, Brigham Young University (1985) · Co-author of Tension: The Energy of Innovation (2015) and Solving for Why (2019)

Views are my own and do not represent Utah Valley University.

Talks

Keynote

What a Finished File Was Never Proving: Rebuilding Assessment for the AI Era

The assessment crisis started before AI. A polished essay, model or deck never proved that a student could do the work alone, and AI made that obvious fast. This keynote shows a design response that doesn’t rely on detection or bans: a fading scaffold, questions anchored to each student’s own work, live defense, and a gradebook that puts the points where independent performance is observed.

Format: keynote, 45–60 min · in person or virtual
For: faculty, deans and provosts, teaching-and-learning centers, education conferences

You’ll leave able to:

  • Explain why grading the polished file now means grading the AI
  • Sort your graded work on the 2×2 grid: with or without assistance, product or observable competence
  • Start three design moves on Monday, with no platform required

Workshop

AI On, AI Off and AI Proof Assessments: A Hands-On Course Redesign Workshop

Bring one real assignment and leave with a redesigned version. Participants map their gradebook on the 2×2 grid, then rebuild one assignment in pairs with a fading scaffold (full, shared, light, AI off), anchored questions and a live defense where the revision carries the points. A short LLM council exercise shows why agreement between models is not evidence.

Format: 90 min to half day · in person or virtual
For: faculty, instructional designers, faculty-development programs

You’ll leave able to:

  • Map your current gradebook onto the 2×2 grid and set a target point split
  • Redesign one assignment with an AI-off checkpoint and anchored questions
  • Run a blind-score, live-defense, revise cycle in your own class

Business and finance education

No Textbook, AI on Every Assignment: How I Teach Graduate Financial Management

Teams work one real company all term from its actual filings and report to the CFO, the CEO and the board. Seventy-seven concepts, each tied to a course outcome, replace the textbook. Students build models by prompt and then fix what the AI got wrong, and two-page memos are defended live with AI off. I walk through what worked, what didn’t, and how to adapt each move without my platform.

Format: session or keynote, 45–60 min · workshop, 90 min · in person or virtual
For: business and finance faculty, business-school leaders, finance-education conferences

You’ll leave able to:

  • Design a backward-mapped course arc anchored in real filings
  • Use LLM councils to teach source discipline and judgment
  • Shift grade weight toward live, AI-off demonstrations of mastery

Innovation and identity · from the Utah HIMSS 2026 keynote

Tension Doesn’t Determine the Outcome. Identity Does.

Proven innovations stall, and the obstacle is rarely the technology. It’s the structure and identity of the organization. Pressure is neutral: organizations respond as Victims who cut and blame, Reformers who optimize inside the system, or Architects who redesign it. This talk covers the three ways to change a system (reform it from inside, regulate it from outside, or route around it) and why unsanctioned AI use is a signal, not just a threat.

Format: keynote, 45–60 min · leadership session · in person or virtual
For: healthcare, education and corporate leaders; innovation and strategy teams

You’ll leave able to:

  • Diagnose whether a stalled innovation is a technology, training or architecture problem
  • Choose among reform, regulate and route-around strategies for your own system
  • Reframe shadow AI as signal intelligence about unmet needs

Formats

Keynote

45–60 min, built around your audience and theme.

Workshop

90 min to half day. Everyone leaves with a redesigned assignment.

Faculty development

A single session or a multi-session series with a redesign sprint.

Virtual

Any format, live online.

See Chris on stage

From Chris’s Utah HIMSS 2026 keynote, Ogden, April 24, 2026.

Watch the full Utah HIMSS 2026 keynote on YouTube (29 min)

Nothing loads from YouTube until you press play.

What organizers say

“Chris brings rare clarity to the architecture problem at the heart of digital health — why proven innovations stall and which organizations capture the upside. His UHIMSS keynote combined the strategic depth of a global healthcare consultant with the operational candor of someone actually running a digital health company, and his patient-first reframe — owning the space between clinical encounters — is the most actionable strategic insight I’ve brought back to our team in years. Our members walked out with work to do, and the structure to do it well.”

Kade Haviland, Board Member, Utah HIMSS Chapter

Recent stages

  • Upcoming · October 30, 2026: How I Teach FIN 6150, Woodbury School of Business, Utah Valley University
  • October 1, 2026: Measured Against MIT, Session 3.4, Finance Education Association (FEA) 2026 · more below
  • April 24, 2026: The 17-Year Gap: Innovation in the Shadow of OBBBA, keynote, Utah HIMSS 2026, Ogden

Past talks and resources

October 1, 2026 · FEA 2026, Session 3.4

Measured Against MIT

What MIT’s AI committee recommendations look like in live finance courses, scored recommendation by recommendation.

Slides by request: chris@crucibl.tech

The MIT scoreboard

Where my courses stand against MIT’s recommendations

MIT’s AI committee made 26 recommendations, as I count them. MIT does not number them and says plainly that they are “not a checklist,” so the count and the split are mine. Fifteen of the 26 can be acted on by an instructor or a department, and the other eleven need the provost.

Against those fifteen, as of Fall 2026:

Scored by Chris Wasden against his own courses, Fall 2026. Source: MIT Ad Hoc Committee on AI Use in Teaching, Learning and Research Training, final report, §3.
StatusCountWhat it means
Done8Running in live course sections this term.
Partly5The instrument exists. The measure does not.
Not yet2Both are department-level in MIT’s own framing.

MIT’s report on AI and education

Research

The design of the AI tutors and the way their help fades is specified in a manuscript currently under peer review.

A within-section randomized trial, modeled on Kestin et al. (2025), is planned for Fall 2027: the same instructor teaching both arms, with external co-authors. If your institution wants to run a parallel arm, get in touch.

The studies behind the method

My own course numbers are illustrative. These studies do the work.

  • Koedinger et al. (2015, 2016). Carnegie Mellon’s Open Learning Initiative: the “doer effect,” where practice predicts learning far more than reading. It is also the most overstated finding in the field. The often-quoted “6×” is a ratio of regression coefficients on observational data. It is not six times the learning, not an effect size, and not a randomized result.
  • Kestin et al. (2025). Scientific Reports 15:17458. A Harvard physics randomized trial of an AI tutor, N = 194: gains of 0.73 to 1.3 standard deviations.
  • Kulik & Fletcher (2016). A meta-analysis of 50 evaluations of intelligent tutoring systems: median effect 0.66 standard deviations.
  • Bastani et al. (2025). PNAS. A randomized trial in Turkish high-school mathematics: unguarded AI access harmed learning.
  • Lepine et al. (2025). arXiv. 34 finance professionals, 1,178 sessions: evidence on cognitive load.

Kestin had 194 students and a control group. My courses have neither a control group nor that design yet, which is why the 2027 trial matters.

What attendees leave with

The Monday Moves worksheet

Three changes to make next week, plus the blank 2×2 grid. Free at crucibl.tech/monday.

AI-Proof

Redesign one of their own assignments, with free redesigns every month. Try AI-Proof free

Departments can also rebuild syllabi with AI-Proof Syllabus (10-pack $349).

Prices exclude sales tax. Purchases open Friday, October 23.

A clear path

A framework they can use the next day, not a pitch.

Stay in touch

Two ways to keep up

Every Tuesday · free

AI-Proof Weekly

One redesigned assignment, one technique and one tool each week, for K-12 teachers and college faculty. Practical things you can do on Monday.

Subscribe to AI-Proof Weekly

You sign up on the AI-Proof home page. Double opt-in, unsubscribe anytime.

Coming January 2027

The Architect’s Brief

A monthly deep dive for higher-ed faculty and leaders: what I am doing with my own courses, why, and how, so you can rebuild yours. Published on my Substack.

Follow on Substack →

Bring Chris to your event

Tell me about your audience and I’ll reply within 2 business days.

Prefer email? chris@crucibl.tech · also reachable at chris.wasden@uvu.edu

For organizers