AI Capabilities Update

What’s Changed Since Fall 2025

Mitch Gerhardt

IDEEAS Lab

2026-05-19

Agenda

10:00–10:45 | Redefining Competencies and Revising Learning Objectives

  1. AI Literacy Foundations
  2. The Evolving AI Landscape
  3. AI Capabilities by Discipline
  4. What Accreditation and Policy Bodies Are Saying
  5. Competency Conversation
  6. Workshop Activities

AI Literacy Foundations

Expanding on Workshop 1

In Workshop 1, we discussed the notion of “AI literacy” from Almatrafi et al.’s (2024) systematic review of AI literacy conceptualization, constructs, and implementation and assessment efforts (2019–2023). They identified six core constructs of AI literacy:

Construct Description
Recognize Knowing AI is present and active in a technological system
Know & Understand Comprehending how AI is functions “under the hood”
Use & Apply Being able to use AI tools effectively, appropriately, and purposefully
Evaluate Being able to critically evaluate the outputs of AI tools and systems
Create Being able to create with AI tools and systems
Ethically Navigate Being aware of the ethical implications of AI tools and systems, and being able to make informed decisions about their use

How have AI capabilities shifted the demands on each of these for us and our students?

Almatrafi, O., Johri, A., & Lee, H. (2024). A systematic review of AI literacy conceptualization, constructs, and implementation and assessment efforts (2019–2023). Computers and Education Open, 6, 100173. https://doi.org/10.1016/j.caeo.2024.100173

The Evolving AI Landscape

A Language Model (LM) and a Prompt

User
types a prompt
Language Model
knowledge: training data only
context window
Response
generated text
Q
✓

Connecting to a Knowledge Base

User
asks a question
Language Model
context window (fills with retrieved docs)
Response
grounded answer
Retriever
semantic search
Knowledge Base
documents · papers · data · embeddings
▲ NEW
Q
doc
doc
doc
✓

Giving the LM Tools

User
sends request
Language Model
decides when to call tools
Response
synthesized
Database
SQL · structured data
External API
live data · web services
Code Runner
execute · compute · verify
▲ NEW
▲ NEW
▲ NEW
Q
fn()
fn()
✓
✓
✓

Coordinating Multiple LM “Agents”

User Task
complex goal
Orchestrator Agent
decomposes · assigns · synthesizes
Agent A
Research & Retrieval
Agent B
Code & Computation
Agent C
Review & Critique
KB · Web search
Code runner · DB
Eval · Scoring
Final Output
synthesized · verified
?
?
?
↑
↑
↑
task
▶
▶
▶
✓
✓
✓
✓

AI Capabilities by Discipline

Civil / Structural

Civil / Structural

Generate and iterate structural load models, interface directly with FEA tools from natural language prompts

Autodesk Assistant - https://www.autodesk.com/solutions/autodesk-ai/autodesk-assistant#item2

Mechanical

Mechanical

Produce CAD specifications from natural language descriptions, optimize thermal and fluid designs iteratively with AI agents

SOLIDWORKS - https://www.solidworks.com/solution/how-ai-is-augmenting-cad-tools-better-product-design

Electrical

Electrical

Debug circuit designs from schematic context, and generate VHDL/Verilog from natural language specifications

Cadence Allegro X AI - https://www.ema-eda.com/products/cadence-allegro/allegro-x-ai-overview/

Software

Software

Autonomous code generation, test writing, and refactoring across full repositories, transforming the software industry

Reeves, B. N., Prather, J., Denny, P., Leinonen, J., MacNeil, S., Luxton-Reilly, A., Nicolajsen, S. M., & Brabrand, C. (2025). Prompts First, Precision Later: Reviving the Vision of Natural Language Programming for Computing Education. Proceedings of the 25th Koli Calling International Conference on Computing Education Research, Koli Calling ’25, 1–8. https://doi.org/10.1145/3769994.3770039 · Vardi, M. Y. (2025, October 14). Computing Is Indeed a Discipline in Crisis – Communications of the ACM. Communications of the ACM. https://cacm.acm.org/opinion/computing-is-indeed-a-discipline-in-crisis/ · Denny, P., Prather, J., Becker, B. A., Finnie-Ansley, J., Hellas, A., Leinonen, J., Luxton-Reilly, A., Reeves, B. N., Santos, E. A., & Sarsa, S. (2023). Computing Education in the Era of Generative AI (arXiv:2306.02608). arXiv. https://doi.org/10.48550/arXiv.2306.02608

Chemical

Chemical

Predict reaction pathways, suggest synthesis routes by reasoning over the published literature, and generate molecular structures with desired properties

Q8W3K0: A potential plant disease resistance protein. Mean pLDDT 82.24. Predicted by AlphaFold via https://alphafold.ebi.ac.uk/entry/Q8W3K0

Biomedical

Biomedical

Interpret imaging outputs at expert-level accuracy and generate structured clinical documentation from voice

Azmed - https://www.azmed.co/news-post/ai-powered-chronology-helps-radiologists-separate-acute-from-chronic-fractures

The Pattern

AI is not replacing engineers, but it is operating in the spaces where junior engineers, graduate students, and methodical routine work used to live

Students Early Career Training Absorbed by AI Professionals Strong expertise to leverage AI Rising expertise expectation as AI takes over routine tasks the Expertise Gap Limited entry-level work to bridge it
Digital Education Council. (2025). AI Skills Opportunity Map. https://www.digitaleducationcouncil.com/ressource-library-items/ai-skills-opportunity-map · OECD. (2023). Bridging the AI Skills Gap. https://doi.org/10.1787/66d0702e-en · Brynjolfsson, E., Chandar, B., & Chen, R. (2025). Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence. Stanford Digital Economy Lab. · Cui, Z. et al. (2025). The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers. SSRN 4945566. https://doi.org/10.2139/ssrn.4945566 · Martin, M. (2025). Educated but unemployed, a rising reality for college grads. Oxford Economics. · Thompson, D. (2025, Oct 2). The Evidence That AI Is Destroying Jobs For Young People Just Got Stronger. derekthompson.org

AI Literacy Is Individual + Collective, Technical + Social

“Is this student AI literate?” → “Is this learning environment building AI-capable practitioners?”

Individual Collective
Technical Personal use, prompting, code generation, model eval Team workflows, shared tools, institutional infrastructure
Social Professional judgment, accountability, authorship Field norms, disciplinary ethics, regulatory frameworks, culture

AI competence is not a property of a person alone, but emerges from how individuals, teams, institutions, and fields negotiate AI’s role. Innovation means knowing when to integrate and when to reject. Both AI adoption and rejection are active, informed, professional choices.

What Accreditation and Policy Bodies Are Saying

ABET

  • Will not prescribe whether or how programs use AI
  • Holds programs fully accountable for student outcomes, academic integrity, and ethical standards
  • Student Outcome (c) — “ability to communicate effectively” — now includes AI-mediated communication

New AI/ML Program Criteria

Graduates must “apply AI theories, models, and techniques to design and implement AI-based solutions”

Accreditation doesn’t mandate AI integration, but it asks: “Can your graduates meet outcomes in a world where AI is present?”

ABET. (December 2025). Updated program criteria and AI/ML accreditation standards.

UNESCO

AI Competency Frameworks for Teachers and Students (September 2024)

For Teachers — 15 Competencies

5 dimensions: Human-centred Mindset · AI Ethics · AI Foundations · AI Pedagogy · AI for Professional Learning

For Students — 12 Competencies

4 dimensions including AI system design, at three levels:

Understand · Apply · Create

Emphasis: human agency, human rights, sustainability, with AI supporting intellectual development, not replacing it

UNESCO. (2024). AI Competency Frameworks for Teachers and Students. https://www.unesco.org/en/digital-education/ai-future-learning

Brazil

MEC Framework (February 2026)

Learning with AI and learning about AI — both required

  • Teachers as pedagogical designers, not implementers
  • HE focus: AI skills, ethics committees, research fundamentals
  • Challenge 6: protecting active learning and academic authorship from automation

CNPq — Research Integrity Policy

Explicit disclosure required for AI use in all funded research

National AI Landscape

PBIA 2024–2028: Ministry of Science, Technology, and Innovation allocates R$ 23.03B (USD $4B) allocated across training and capacity building in AI

Bill 2338/2023: Passed Senate Dec 2024 — risk-based AI regulation aligned with EU AI Act

MEC. (February 2026). Framework for the Development and Responsible Use of AI in Education. · CNPq. Policy on Integrity in Scientific Activity. · PBIA 2024–2028. Brazilian Artificial Intelligence Plan.

Competency Conversation

What This Creates: A More Complex Competency Landscape

For engineering educators, this means three things are true simultaneously:

Your graduates will work with AI systems and the sophistication of those systems is increasing rapidly

Your graduates will be evaluated and regulated by bodies (ABET, CONFEA/CREA, employers) whose own frameworks are still catching up to AI

Your course is where the negotiation happens between foundational knowledge, professional competence, and AI-era capability

The Open Question

At your institution, in your discipline, and with your students

What does it mean to know your field?

What does it mean to do your field?

And how, if at all, does AI change those answers?

No “correct” answer, but it shapes how you design learning experiences, and how your students understand their future roles as practitioners in the field

Workshop Activities

Workshop: Your Syllabus in the AI Era

You submitted your syllabus. Now we use it.

Step 1 — Map (~8 min)

Using the AI-era competency matrix (individual/collective × technical/social), identify where your existing learning objectives already address AI-era competency and where they don’t

Step 2 — Identify (~7 min)

Which objectives need revision? Which need to be added? Which should be reframed rather than replaced?

Step 3 — Draft (~12 min)

Write 3–5 revised or new learning objectives that explicitly incorporate AI literacy without abandoning core disciplinary competence in your course

Step 4 — Stars & Steps Feedback (~8 min)

Exchange drafts with a disciplinary peer (same or adjacent field)

Stars & Steps: A Protocol for Peer Feedback

We’re going to be using this simple protocol throughout the workshop:

Stars — What’s working

Identify 2 specific strengths in your partner’s revised objectives:

  • What is precise, actionable, or well-framed?
  • What reflects real understanding of the AI challenge in this discipline?

Steps — What could go further

Offer 1–2 specific, constructive suggestions:

  • Where is an objective still too vague or too broad?
  • Is there a missing dimension (individual/collective, technical/social)?
  • Does the objective actually require AI literacy, or could it be met without engaging AI at all?

Ground rules: Stars always first, with steps being directional, not prescriptive. We’re pointing, not rewriting · 3–4 minutes per partner

Thank You

Questions?

Mitch Gerhardt

mitchg@vt.edu

github.com/mitchg10

linkedin.com/in/mitchell-gerhardt