AI + OER Toolkit
Everything you need to create open educational resources with AI responsibly: a values check, a quality rubric, licensing guidance, an accessibility checklist, and a plan for bringing faculty along.
OER Values Compass
Evaluate every AI-powered OER workflow against these five values.
Access
Can a student reach this content without a paywall, a login wall, or a device they cannot afford?
Affordability
Does this reduce cost for students, faculty, or your institution?
Quality
Is the content accurate, current, and pedagogically sound?
Sustainability
Can your institution maintain this after the grant cycle ends, without a vendor relationship?
Openness
Can it be shared, adapted, and reused under a clear license?
Critical Questions in AI and Open Education
These are the questions your faculty will ask. They are the right questions. Here is what we know, what we do not know, and what you can do.
1 Training Data and Copyright
"What about the copyright issues with AI training data?"
AI models were trained on enormous amounts of text, and there is real, unresolved litigation about how much was used with permission. The legal landscape is still being mapped. No one can tell you today with certainty where the courts will land.
What you can do
Use AI to generate original content from openly licensed sources, not to reproduce or summarize copyrighted texts. The OER prompts in our Higher Education Prompt Library are built to generate new content from the structure of openly licensed material. See the Licensing Guide below for the current legal landscape.
2 Privacy and Student Data
"What data are we putting into these tools?"
Anything you type into a general-purpose AI tool is processed by a third party. Putting student work, grades, or personally identifiable information into these tools raises FERPA concerns and, on most campuses, requires a conversation with IT and general counsel.
What you can do
Do not enter student data into AI tools. Use only openly licensed source material as inputs. Our OER prompts use OpenStax and OER repository content as inputs. Check your institution's AI acceptable use policy before recommending any tool to faculty.
3 Environmental Cost
"What is the environmental cost of running these models at scale?"
Training and running AI models requires significant energy. The numbers are not small, and they are growing. This is a real cost that the field has not fully accounted for.
What you can do
Use these tools deliberately. A single well-engineered prompt that produces usable output is far less wasteful than 20 poorly written prompts. Well-structured prompts, like the ones in our Higher Education Prompt Library, are partly about quality and partly about reducing wasted computation.
4 Content Accuracy and Completeness
"How do I trust what AI produces?"
AI tools confidently produce wrong answers. They produce outputs that miss important content, contain factual errors, or are pedagogically thin. None of this disqualifies AI from OER work. It means everything AI generates requires human review by someone who knows the subject matter.
What you can do
Run every piece of AI-generated content through the OER-AI Evaluation Rubric before it reaches students. Use the Accessibility Audit prompt to catch structural issues. Treat AI output as a first draft, not a finished product.
OER-AI Evaluation Rubric
Apply this framework to every piece of AI-generated content before using it with students.
1 Accuracy
Factual claims are correct, current, and verifiable against the source discipline.
What to look for
- Claims match the openly licensed source material
- Statistics and dates are verifiable
- Definitions align with the discipline's accepted usage
Red flags
- Confident-sounding but unverifiable claims
- Outdated statistics or superseded theories
- Plausible-sounding fabrications (hallucinations)
2 Alignment
Every question, task, and activity maps directly to stated learning objectives.
What to look for
- Each assessment item tags a specific LO
- Bloom's level of the question matches the LO verb
- Content stays within the scope of the specified section
Red flags
- Questions that test content not covered in the LOs
- All recall, no application or analysis
- Scope creep into adjacent chapters or topics
3 Content Accuracy & Completeness
Content covers the breadth of the topic without significant gaps, omissions, or narrow framing.
What to look for
- Multiple perspectives within the discipline represented
- Key concepts from the source material not omitted
- Examples are current, varied, and broadly applicable
Red flags
- Important sub-topics silently dropped
- Examples drawn from a narrow range of contexts
- Oversimplified treatment of complex concepts
4 Accessibility
Content meets WCAG 2.1 AA standards and is usable by all students, including those using assistive technology.
What to look for
- Reading level appropriate for the audience
- Clear heading hierarchy (H1 > H2 > H3)
- Alt text for images, text alternatives for data
- No meaning conveyed by color alone
Red flags
- Dense academic jargon without definitions
- Skipped heading levels or no heading structure
- Charts or tables with no text summary
5 Licensing Clarity
Content has a clear, appropriate Creative Commons license and proper attribution for all source material.
What to look for
- CC license type stated clearly
- TASL attribution for source material included
- License is compatible with source license
Red flags
- No license stated ("all rights reserved" by default)
- Incompatible licenses mixed without resolution
- Attribution missing or incomplete
Creative Commons Licensing Guide
Quick reference for licensing AI-generated OER content.
Who Owns AI-Generated Content?
What the law says
Purely AI-generated output is not copyrightable (US Copyright Office Part 2 Report, Jan. 2025; Thaler v. Perlmutter, D.C. Cir. 2025, cert. denied Mar. 2026). Detailed prompts alone (even when iterated extensively) do not constitute authorship.
Copyright can attach when you:
(a) Creatively select, coordinate, and arrange AI outputs
(b) Substantively modify or edit those outputs
(c) Incorporate your own expressive inputs that remain perceptible in the final work
What to do for your OER
1. Apply your CC license to the human-authored elements and your original selection and arrangement.
2. Mark substantially AI-generated portions as CC0 / public domain (Creative Commons' own recommendation).
3. Document your prompts, edits, and creative decisions.
4. Add a transparency statement disclosing AI use.
Human-curated, human-edited AI content is on solid ground for open distribution. Do not assume strong copyright control over the AI-generated portions themselves.
CC License Options at a Glance
| License | Share | Adapt | Commercial Use | Same License Required | Best For |
|---|---|---|---|---|---|
| CC0 | Yes | Yes | Yes | No | Public domain. CC-recommended for substantially AI-generated content. |
| CC BY | Yes | Yes | Yes | No | Maximum reuse with attribution. OpenStax default. |
| CC BY-SA | Yes | Yes | Yes | Yes | Keeping derivatives open (Wikipedia model). |
| CC BY-NC | Yes | Yes | No | No | Sharing but preventing commercial sale. |
| CC BY-NC-SA | Yes | Yes | No | Yes | Maximum protection while still sharing. |
OpenStax + Your AI Content
OpenStax textbooks are licensed CC BY 4.0. The OER Builder and your manual prompts reference structural elements only: chapter titles, section names, and learning objectives. No textbook text is reproduced, so your output does not infringe the OpenStax source.
Copyright protection applies to your human-authored contributions (your edits, your selection and arrangement, your expressive additions) not to the raw AI-generated text itself. Use CC0 for substantially unedited AI-generated portions. Attribution to the OpenStax source is required in all cases.
Which License Should I Pick?
1. How much did you edit? Substantial edits, selection, and arrangement: apply a CC license. Largely unedited AI output: apply CC0.
2. Sharing publicly? If yes, you need a license. LMS-only is optional but good practice.
3. Want others to adapt? All CC licenses allow it. SA requires derivatives use the same license.
4. Allow commercial use? Yes = CC BY or CC BY-SA. No = add the NC clause.
For human-authored OER, CC BY is the community standard. For AI-generated portions, CC0 is the safest choice.
TASL Attribution Format: Title of the work, Author, Source (URL), License (with link to deed). Example: "Introduction to Sociology 3e" by OpenStax, https://openstax.org/details/books/introduction-sociology-3e, CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
Add a Transparency Statement to Your OER. Disclose what AI tools were used, how they were used (prompting, editing, arranging), and the extent of human modification. This aligns with US Copyright Office registration guidance, OER community best practices, and protects you if you ever register the work. Example: "This module was drafted with [AI tool name] using structured prompts based on OpenStax learning objectives. All output was reviewed, edited, and arranged by [Your Name]. AI-generated content was checked for accuracy against the source material."
Legal Disclaimer. Copyright law around AI-generated content is actively evolving. The key case Allen v. Perlmutter (D. Colo.) was fully briefed and still awaiting a ruling at our last review (October 2026) and could shift the threshold for AI authorship. International law differs significantly (the UK and China have reached different conclusions than the US). This guide provides general educational information, not legal advice. Consult your institution's legal counsel or OER librarian before making licensing decisions for formal publication or grant compliance. Laws and agency guidance may change after this page was last updated.
Last reviewed: October 2026
Accessibility Checking Guide
Run this checklist on every piece of AI-generated content before it reaches students.
Quick Checklist
- Reading level is appropriate (grades 13-14 for community college)
- Specialized terms defined inline or in a glossary
- Heading hierarchy is correct (H1 > H2 > H3, no skipped levels)
- All images have descriptive alt text
- Tables have text summaries or alternative descriptions
- No meaning conveyed by color alone
- Instructions and questions are unambiguous
- Numbered lists for sequential items, bullets for unordered
- If assessment, multiple ways to demonstrate understanding
- Ran a text search on distinctive phrases to check for training-data regurgitation before publishing
Free Checking Tools
Checks web pages for WCAG compliance. Free browser extension.
Checks reading level and sentence complexity. Paste your content in.
Tests color contrast ratios against WCAG AA and AAA standards.
Right-click > Inspect > Lighthouse tab > check "Accessibility" > Generate report. Built into Chrome.
The Accessibility Audit Prompt in the Prompts section above is another option: paste your AI-generated content and ask any AI tool to audit it against WCAG 2.1 AA.
Tool Links Directory
Free tools for building, checking, and sharing open course content.
General AI Tools
Eclatech Tools (Free for Educators)
OER Sources
Accessibility Tools
Faculty Adoption Models
These four models are a progression, not a pick-one menu. Start at the Invitation. Move through them as faculty interest grows.
The Invitation
Awareness
30 minutes
30-minute lunch-and-learn at your library or center for teaching. Show 2-3 AI-powered OER workflows live. No commitment required. Participants watch, ask questions, and leave.
Best for: Campuses where faculty have not seen AI applied to OER.
Faculty walk away with
Resource page link and prompt templates. No obligation.
The Proof of Concept
Interest
~1 hr (yours) + 15 min (theirs)
Build a complete module yourself using a faculty member's course materials and openly licensed sources. Show them the finished product. They never touch a tool. Works especially well after an Invitation session.
Best for: Faculty who need to see results before engaging.
Faculty walk away with
A ready-to-upload module they did not have to build. Proof that this works for their course.
The Co-Build
Trial
30 minutes
Sit down with one faculty member. You drive the prompt. They watch. Together, generate one content piece from their course. They edit it with you before you leave. Best after they have seen a proof of concept.
Best for: Skeptical or time-constrained faculty who have already seen results.
Faculty walk away with
One usable content piece they co-created, plus the prompt template to do it again.
The Department Pilot
Integration & Scaling
One semester + monthly check-ins
One department, one semester, one course. Provide a workflow guide and monthly check-ins. Collect outcomes: time saved, student feedback, content quality ratings. The pilot becomes the evidence base for broader adoption.
Best for: Departments with an OER champion. Programs needing documented outcomes.
Faculty walk away with
Documented results they can present to their dean, provost, or grant officer. A repeatable process for the next course.
All four models share one principle: start with one workflow that solves one problem the faculty member has right now. Win once. Scale from there.
Action Plan: Next 30 Days
Fill this out, then copy or print it. This is what you bring back to your campus or department.
Want hands-on AI + OER training for your faculty?
Eclatech Solutions delivers practical AI training for educators, departments, and institutions.
Plan Your Training