We value respecting privacy, consent, and intellectual property rights. We’ve outlined below our current process and how we intend to handle copyright auditing, student privacy, and compiling course material below.
Coursetexts Process (July 2025)
We currently publish text-only, original content created by the professor of the course. We define “original content” as materials created by the consenting professor for the course (e.g., lecture notes, syllabi, and assignment prompts). We omit materials created by co-instructors, guest lecturers, students, or prior course staff unless we’ve secured their explicit consent.
We screen original content for third-party images and personally identifiable information. Our process for publishing a course is as follows:
- We individually email professors asking permission to publish their course content.
- We create a password-protected preview of how their course will display on our website. In order to create the preview, we take their course content with their consent, replace all copyrighted/third party content with external links, and blur non-professors faces in videos.
- We attach a letter informing the professor of the implications of open sourcing their course under the Creative Commons license.
- Upon approval, we publish the course by removing password protection and linking it from our homepage.
Our Priorities
Consent: We publish with the professor's explicit consent. To inform professors of the implications of open sourcing their course, we send them this letter. We redact any course information that is not original content by the professors who have consented, such as student material or material created by other instructors who did not consent to sharing their content.
Copyright: We currently omit any documents containing copyrighted works owned by third parties whose consent we have not obtained.
Privacy: In compliance with FERPA, we redact all student personally identifiable information (PII) from course materials. In text content, this usually involves redacting names and emails.
Compiling Course Material: We manually compile course material by asking professors to share email us their course content or individually downloading files hosted online.
Our Vision (Q3 2025)
Copyright
We are currently building human-in-the-loop AI pipelines to review copyrighted content. Our copyright guidelines are below:
- Always attribute media when possible. We attribute a source regardless of whether we claim fair use. Below are methods to identify the attribution source:
- Watch the lecture recording, if it exists, to see if the instructor mentions the paper title or paper authors. If it seems like a paper or reading of some kind, we can check the syllabus to see which readings were assigned for this particular lecture, and look through all the readings to try to find the media.
- Reverse image search (type “reverse image search” on google and try using a few different options for reverse image search)
- If the media has text, Google the text to find the source.
- In the worst case: write “source unknown” and we will ask the professor about all unknown sources in our next communication with them.
- For each piece of external media, we will do one of three things:
- Claim fair use (action: do nothing)
- Get explicit permission from the source to use it (action: email the source)
- Mark the piece of media as something we have to recreate or replace with a substitute (action: mark to substitute later). If we cannot find a substitute, we will recreate the piece of media ourselves. For example, if it’s an infographic or illustration, we will hire a graphic designer and make our own version.
- When should we claim fair use?
- For most pieces of media, we will claim fair use. We cannot claim fair use for content unique to a textbook, especially textbooks that are centered on the same topic as the course. For example, 14.54: “International Trade” uses the textbook “International Economics” — our course and the textbook are too similar in their purpose and target audience to claim fair use. For any piece of media from the textbook, we will either have to get explicit permission from the publisher, or recreate it.
- When should we ask for permission?
- Don’t bother asking for permission from the source when we expect the answer to be no. We expect the answer to be no in most cases, e.g. newspaper publishers like The London Times. The guideline is to only ask for permission for media from an academic paper and academic-ish organizations like non-profits and maybe some academic paper publishers.
- How do we ask for permission?
- Email both the paper authors, and the paper publisher separately (two emails).
You can find additional detail about our work with MIT SOUL and MIT OCW on copyright and attribution at this link.
Compiling Course Material
We currently aggregate course content manually, so all course notes are downloaded by an individual. The legal and product teams at Canvas have approved automatic course material downloads by using the Canvas API. We are currently seeking permission from individual universities to use the Canvas API.
We are currently working with MIT OCW to develop a video pipeline that can automatically remove student video and audio with state of the art accuracy.