Updates to our terms and new GenAI Policy
Comments
-
@SansaoSeara Hello! Thank you for your questions:
Hoping that my responses can bring you (and others) more clarity around the GenAI Program.
1 - Will or does Udemy intend to create AI-generated courses?
Udemy does not have any plans or intentions of creating AI-generated courses that compete with instructors.
2 - Will this impact the revenue sources we already have, Marketplace, UB, individual plan?
The current revenue models for Marketplace, UB & Personal plan will remain the same.
3 - What are the derivative works that can be generated and how do they generate revenue? revenue?
The derivative works include the output for our GenAI tools (AI Learning Assistant & Learning Paths). The monetization plan of that content is currently being finalized by our leadership team.
4 - Will the marketplace, ads and organic sales continue to exist in their current format?
Yes, the Marketplace, ads and organic sales will continue to exist in their current format. There will be no change.
Thank you again for your thoughtful and respectful questions.
1 -
@Chrystie … Can we get bullet #1 incorporated into the terms by specifically stating that our content will not be used to create AI-generated courses?
1 -
"Udemy does not have any plans or intentions of creating AI-generated courses that compete with instructors."
It is too obvious that the wording of everything is crafted in a way that can always be open to interpretations, even in communications that don't have any legal significance.
I can see a very small course covering a very specific topic, targeting a very specific audience, derived from one of my courses with the explanation that it does not compete with mine, because mine is much more comprehensive and broad. It's easy to imagine many more scenarios.
I don't mean to be negative or argumentative, but I don't like to be tricked. If there is nothing to hide, language should be much more direct and unambiguous.
2 -
@Chrystie what is a learning path and how is that different from a course?
Is a learning path bits of our already existing courses combined and sold to students?
0 -
Hello @ThomasMitchell
We understand that changes like these can be challenging, and appreciate your passion and engagement in the community.
Please know that the GenAI policy was developed with considerable time and effort to ensure it aligns with our AI principles and legal obligations. We considered many factors and made every effort to create a policy that is fair and beneficial for the community as a whole.
At this time, we do not anticipate making any changes to the policy. However, we are committed to continuously reviewing and improving our practices.
0 -
@ElizabethEvans Great question. You can see the Learning Paths in action in this webinar replay.
0 -
@Chrystie That’s unfortunate. Any objective person can only read that as “I know what we’re promising you, but we’re not committing to it legally”.
I recognize that my individual action is but a blip on Udemy’s radar, but I just don’t see myself being able to opt in.
Between the demo that only showed 1 or 2 courses being included in the AI learning path builder, the refusal to legally ensure my content won’t be used to build AI-generated courses, the lack of clarity on how usage of my content by the AI will be tracked, and the lack of clarity over how I’d be compensated, I just can’t find any benefit to me.
I really wish that wasn’t the case. But it is what it is.
Ironically, it’s the performance of the platform and the revenue cuts that have led me here, because there was a time when I earned enough here to accept terms I wasn’t happy with due to FOMO. Those days are gone. I don’t make nearly enough to elicit that FOMO feeling anymore.
4 -
This does make me a little nervous my course represents over five hundred hours of work and a lot of personnel sacrifice and I must admit I don't like the idea of Udemy extracting content from it and republishing without input from me. Further more I don't see if Udemy is realesing new AI genarated courses how they are not compeating with my own content. So far I have not seen anything to show me what these AI courses will look like and exactly they will work.
If a potential student in my case types Blender into the Udemy search engine will the AI Course be displayed in the same column as my course and will they be undercutting my course. If these questions could be answered in a concise manor that leaves no doubt I would be more inclined to stay in, but with so much currently unanswered I will be opting out.
0 -
Maybe it has already been talked about elsewhere, but I have been reading through "Instructor Generative AI Policy", "Terms of Use", and "Instructor Terms" and I couldn't find anything new about making changes and updates to content from courses that have been opted into Udemy's GenAI program.
Are there any new rules instructors need to be aware of before making changes or updates to content in courses included in the GenAI program?
If someone can get back to me soon that would be great because if there are I would like to be able to make all the changes and updates I have planned for my Python course before the GenAI opt out deadline on September 12.
0 -
I'm not a lawyer, but since I worked with European IP rights a lot, I consider two things especially problematic: 1) Udemy not being specific and clear about how our intellectual property is used (and no, Udemy's "We have no plans / we have no intentions" is probably not enough here), 2) offering an opt-out solution instead of an opt-in solution. ChatGPT puts it like this:
"Udemy needs the right to use your content for training AI. These rights must be clearly (!) stated in the contract. If Udemy is using your content for purposes beyond what was originally agreed upon, this could violate copyright law, especially if you did not explicitly consent to a specific use. (…) Under EU law, especially the Unfair Contract Terms Directive (93/13/EEC), terms in a contract must be clear, transparent, and specific. If Udemy’s terms are vague or ambiguous about how your content will be used, this could be problematic because you, as the content creator, may not fully understand or agree to the scope of the use of your intellectual property."And:
"An opt-out mechanism, where Udemy assumes they have the right to use your content unless you explicitly say no, could be legally problematic, depending on the jurisdiction and the specific terms of your agreement with Udemy. In many legal systems, especially under EU law, it is generally expected that for new or expanded uses of IP (like training AI models), the rights holder (you) should have the ability to opt-in rather than being automatically included and then having to opt out."
- Don't know if Udemy checked their approach's alignment with proper EU law, but at first glance, there seem to be a couple of problems.
2 -
@Chrystie
You wrote: We will exclude opted-out content from our GenAI tools by December 31, 2024, and will notify you via email once our AI tools have been updated. During our pre-release testing phases, which run through the end of 2024, your content may be used as we refine these tools. However, we will ensure that it is excluded before any public launch of these AI features.How can Udemy think this is acceptable or even legal to use our content for ANYTHING if we, as content creators, told you we wanted to opt out. I'm not even going to mention forcing everyone to opt in in the first place.
As of today, I opted out of this new dystopian feature, and you, Udemy, are expressly NOT allowed to use any of this content. If you confirm that you will still be using my content, even after I opted out, I will be pulling my courses out of Udemy and taking you to court. Unlike many on your platform, I'm not a small part-time content creator that you can bully around.
You can make all the promises you want that you won't use content to compete with human instructors, but at the end of the day, your policy is written to say that you can.
THIS IS NOT ACCEPTABLE.3 -
Hello @GregReverdiau
Thank you for taking the time to share your concerns with us. I want you to know that we hear you and understand the gravity of your concerns.
The intention behind our pre-release testing phases is to ensure that the AI tools we’re developing are refined and effective before they go public. Your trust is important to us, and we want to be transparent throughout this process.
I encourage you to watch the GenAI Product Preview replay to fully understand the benefits we’re aiming to provide with this program. However, we fully respect your decision if you feel that this program isn’t right for you. If you choose to remain opted out, we’ll ensure your content is excluded before any public launch, and you’ll receive an email confirming this by the end of the year.
We truly value your contributions to Udemy. Please don’t hesitate to reach out if you have further questions or need additional clarification.
0 -
Frankly Christie, there were not too many concrete details in that session either.
0 -
This all sounds a lot like what Adobe does. We use your content to train our models, whether you like it or not. We say A in our communications, say B in our terms of service and eventually do C. It's all too vague. I have nothing against AI, but there need to be clear terms of service and the fact that you use my contentcon without my permission even when I opt out, tellst me enough. You are not trustworthy at all and whenever you can you will exploit the vaguevagueness in the terms of service. Another reason to opt out is the lack of a clear compensation model. That should already be in place. Being vague about that again tells me only one thing, you will avoid fair compensation as long as you can. In fact we should already be compensated since you use our content to train the models without our consent. Because you are intentionally vague I'm going to opt out. As long as you say A and write B in your terms of service you can't be trusted with my content, because as everyone else you will exploit all of this and do the unexpected C. Your terms and compensation are intentionally vague and that means you can't be trusted. It all seems to be written for your benefit and eventually you will find loopholes to use my content without compensating me and directly compecompete with me through a backdoor.
1 -
Just opted out of the program, that was a no-brainer decision after seeing the ambiguous responses by the Udemy staff here.
I'll also send an email to the support, clarifying that using my course material in any form to "train" or "prepare" or "test" the new AI in that little gap between opting out and actually being taken off from the program is absolutely prohibited.
We didn't agree to this when we signed up and Udemy can't just do that without our specific consent, I advise you to do the same if you don't want them to use it.
1 -
Oh this is a lot to take in! The EU are spearheading regulations to incorporate AI in safer ways. It's worth cross checking what they're up to for best practices: https://www.europarl.europa.eu/topics/en/article/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence
0 -
As it appears, the new EU-regulation would imply serious issues for the GenAI project, given that it likely would be categorized within the brackets of:
A. Unacceptable risk: Cognitive behavioural manipulation of people or specific vulnerable groups.
B. Unacceptable risk: Social scoring
C. High risk: Enforced EU registration: Education and vocational training
D. Generative AI, like ChatGPT, will not be classified as high-risk, but will have to comply with transparency requirements and EU copyright law:
- Disclosing that the content was generated by AI
- Designing the model to prevent it from generating illegal content
- Publishing summaries of copyrighted data used for training
The Parliament adopted the Artificial Intelligence Act in March 2024 and the Council followed with its approval in May 2024. It will be fully applicable 24 months after entry into force, but some parts will be applicable sooner:
- The ban of AI systems posing unacceptable risks will apply six months after the entry into force
- Codes of practice will apply nine months after entry into force
- Rules on general-purpose AI systems that need to comply with transparency requirements will apply 12 months after the entry into force
High-risk systems will have more time to comply with the requirements as the obligations concerning them will become applicable 36 months after the entry into force.
But, this was just as predicted earlier.
Regards
1
