1
Cognitive Surrender
What an AI Learning Penalty Might Mean for All of Us
New research shows heavy AI use may erode critical thinking and recall. Where's the line between offloading a task and offloading the thinking that makes us better at our jobs?
The AI Debate: Let's discuss
2
LET'S START HERE
What is something you used AI for in the last week? Use the Zoom chat to share some examples
Writing?
Summarizing something you need to understand?
Build a new process?
Help you comprehend something?
Automate something?
Learn something new?
Do something fun?
Cognitive Surrender: The AI Debate
3
NOT ALL AI IS HARMFUL
Laying the groundwork: our research
Cognitive Surrender: The AI Debate
Our research focused on:
Negative effects of offloading
Potential impacts to intrinsic motivation
Impacts to the developing brain with the writing process
Impacts on critical thinking
4
THE HEADLINE RESEARCH
The Generative AI Learning Penalty
Strömberg, Lei & Wu (2026) — 27,000 students, grades 7–12, China, 2023–2025
c
80%
of students adopted generative AI for schoolwork
50%
of AI-using students fully outsourced homework within 5 months
This study is about schoolwork. But the mechanism it uncovers — a task fully offloaded, a skill never tested — isn’t unique to classrooms.
!
The paper states plainly: generative AI has a substantial negative impact on student learning — with no evidence of a corresponding benefit.
Cognitive Surrender: The AI Debate
c
6x
Learning loss compared to pandemic
5
AN HONEST QUESTION
How do LLMs affect the brain?
Ask around your own team and you'll likely hear a pattern: work that looks finished, was produced fast, but that completely falls apart the moment someone asks a single question. We realize the comprehension just. is. not. there.
?
Discussion prompt
When was the last time someone sent you something they built with Claude code - but didn’t actually review? Or something clearly written by AI…? Daphne called it, “lobing AI slop grenades at each other”...
Cognitive Surrender: The AI Debate
6
A DECLARATION
“We cannot choose to become idiots.” - Dr. Roberto Serrano, Brown University
Cognitive Surrender: The AI Debate
College class example: 96% pass midterm with AI -> 49% fail final without AI
Adults in America more likely to have placed a bet than to have read a book
30% of adults cannot paraphrase or make inferences
7
RETHINKING THE BAR
Our Youngest Workers Are 14% Better Communicators But Suffer With Soft Skills Critical for Working with AI
18%
Below average in critical thinking
17%
Below average in attention to detail
10%
Below average in creative problem solving
A skills test shows what someone can produce with AI. It doesn't show whether they can catch AI being confidently wrong. That takes judgment under ambiguity, the kind old-school analytical logic training was designed to capture.
Significant shifts in the importance of core skills have emerged for organizations. Leadership, social influence, AI and big data, talent management, service orientation, and customer service are all increasing in relevance. However, despite the public discussion on the need for critical thinking and double checking work produced by ai, dependability, attention to detail, and quality control have all decreased in importance.
Cognitive Surrender: The AI Debate
!
8
MOTIVATION SCIENCE
Increased Productivity | Decreased Motivation
Durable motivation comes from within (curiosity, mastery, autonomy). Extrinsic motivation (titles, optics, approval) can crowd it out.
Intrinsic
“I want to get better at this.”
•
Curiosity about the problem itself
•
Satisfaction from figuring it out
•
Growing mastery and judgment over time
Extrinsic
“I want recognition for my achievements”
•
Titles and output metrics
•
Looking capable to managers
•
Avoiding the friction of struggle or failure
Are we optimizing our teams (and ourselves) for optics at the cost of building real capability? (and reducing intrinsic motivation over time…)
Cognitive Surrender: The AI Debate
9
Learning is a result of a feedback loop that breaks with improper use of LLMs
10
Ok - so we spent a lot of time discussing issues. Let’s talk solutions.
Cognitive Surrender: The AI Debate
If you’re willing to join the conversation - come off mute, camara on if willing, and able!
11
GeoffreyWortham.too@gmail.com
(private tutoring: Advanced Math, CIS)
Kelly@forwarddigital.org
Test & Learn Community | Optimizers Conference | Hire me
(UX research, Decision Science, Process, Executive Communication)
WANNA CONNECT?
12
OUR RESEARCH
Cognitive Surrender: The AI Debate
https://buildcognitiveresonance.substack.com/p/the-50-percent-problem
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618
https://www.media.mit.edu/publications/your-brain-on-chatgpt/
https://www.insidehighered.com/news/faculty/learning-assessment/2026/07/08/brown-professor-suspects-most-his-class-used-ai-cheat
https://www.media.mit.edu/articles/ai-s-impact-on-cognitive-ability-mit-study-reveals-more-troubling-data/
https://www.weforum.org/publications/the-future-of-jobs-report-2025/in-full/3-skills-outlook/
13
APPENDIX
Cognitive Surrender: The AI Debate
14
THE CENTRAL TENSION
Novices vs. Experts: Who's Actually at Risk?
Teachers / Experts
Already own the underlying judgment
Use AI to compress time on tasks they've already mastered. Achieving more with less. But, according to a study by MIT, heavy reliance on AI still weakened recall. Expertise may buy a longer runway, but it does not provide immunity.
Students / Novices
Still building the underlying judgment
The efforts that build expertise: struggling, failure, revising. Asking questions of mentors… This is exactly what's being automated, often before a novice logs a single rep. Novices are losing the mentors (AI cannot be a teacher!) No struggle here…
?
If junior folks never do the hard work, never have to sweat. If we’re using AI as our jr tech…who trains the next generation of experts?
Cognitive Surrender: The AI Debate
15
IF WE HAVE TIME
Is AI the Real Story of the Layoffs?
100,000+
US job cuts cited AI through mid-2026; nearly double all of 2025
90%
of executives say AI had zero employment impact at their own company (NBER)
0
connection between people who are achieving ROI and layoffs
“AI-washing”: companies overstating AI's role in cuts they'd likely have made anyway. AI reads to investors as ambition, not admission.
?
If leaders are cutting people in AI's name, but the remaining workforce is offloading its own thinking, are we cutting capacity we'll need to rebuild later? And if we are losing the mentorship pipeline…are we setting ourselves up for a huge failure?
Cognitive Surrender: The AI Debate