I'm just curious how many of you have begun experimenting with AI tools (ChatGPT, GPT3 API, AI platforms, etc.) to refine your design and development of e-learning modules that you create in Rise and Storyline. Any good time-savers or other best practices that you're using?
Is using AI to edit stock images allowed under the stock photo licenses? Is there an amount of altering that would qualify it as a new image that is not longer copyrighted?
Lee-Ann, you raise an interesting point. The answer is going to depend on the license terms under which you obtained the photo.
I am not a lawyer, but my understanding is that most stock photo companies allow you to modify (non-editorial) royalty-free images they license to you, subject to certain restrictions (e.g., not making it seem like the person in the photo is supporting a particular political position, or creating "harmful" content).
In practical terms, it would seem like it makes no real difference if I edit the photo or the AI edits the photo, but I was surprised to see that at least one large stock photo company has a clause in their license agreement specifically addressing AI use.
Unfortunately, it's written just vaguely enough that I can't quite tell what it means. It *seems* (to my untrained, non-lawyer eyes) to be saying you can't use their images to TRAIN an AI. I'm not really sure if lawyers would consider using the image in a Midjourney prompt as training the AI (I wouldn't consider it to be such a use, but I'm not a lawyer). There's also the question of whether letting the AI see it would be considered redistribution of the stock photo (something that's usually not allowed).
Taking a super-strict/restrictive stance that disallows any use of stock photos with AI tools would ultimately be counterproductive to the stock photo companies, in my opinion. What would happen is that this would create an incentive to just use the AI tools without the stock photo input. Once the AI tools are capable enough that we can get the specific image we want for our use case from the AI, why would we need the stock photo company at all? But this very fear is likely to cause overreaction from the stock photo companies and an attempt to maintain the status quo by disallowing AI use in the short term.
There is also an understandable fear among photographers and artists that supply photos, illustrations, vector drawings, and other images to the stock photo companies that this source of income is under threat from AI art tools. There is a movement among artists to require AI companies to obtain the artists' permission before using any of the artists’ images to train an AI. If enough artists band together and are able to establish this as a new rule, then it will become embedded in license agreements and laws. In the meantime, stock photo companies may feel an obligation to the artists who feed images into their libraries to take stances such as “no AI use” that the artists want, even if in the long term, such a stance is counterproductive to the company.
All of which is just a long-winded way of saying that it's going to be interesting times for a while until all this settles out.
Lee-Ann - that's a really important point you make. This whole area is currently a bit of a minefield.
Legal concerns have been raised by the major stock image providers and there have already been discussions between Open AI and the image providers, as It seems likely that the 650 million or so images used for training AI text to graphics engines does include some copyright material - even though Open AI says it is taking, and has taken steps, to avoid this. But even if this issue is resolved (and it is likely to be so amicably), it still leaves the question you raise.
Personally, we are not going risk it. I think taking an image purchased as royalty-free under the terms of a usage licence does allow some latitude to modify that image before you use it in a project - I mean Shutterstock, itself, allows you to filter it, change colour, crop it etc. when you download it! But using it as input to generate an entirely new image - or set of images using AI may well be in breach of the terms of that licence.
My assertion that the dispute between image providers and Open AI will be resolved amicably is born out by the recent partnership announced by Shutterstock and DALL-E. Shutterstock will now be offering an AI generated image service! What the cost per image will be I don't know but looking at my current spend with Shutterstock, I doubt it will be cheap! But this seems the way it will go...
Re-reading my reply, I should perhaps add that most of us buy our stock images under a "single-use" licence as it is much cheaper. i.e. if you purchase images using a Shutterstock monthly "plan" and download an image under that plan you are purchasing a licence to use that image for a single project.
We are always at pains to point this out to clients when we supply source files (which is most of the time). They can modify the course and use selected images from the course as thumbnails to promote that course BUT they are in breach of copyright if they then take an image from one of our courses and use it in a different course without paying a new licence fee
To purchase an image "royalty-free" is generally way more costly and we do it only when the client requires that image for multiple projects.
I have recently been very interested in AI. Five days ago, I really started to delve into the topic and think about how it could help with Storyline. All the ideas listed here are great, but AI can do much more with course creation. For example, you can currently use Decktopus, Beautiful.ai, or Slidesgo to create a custom slide deck of your topic all by typing into a box your parameters like colors, look and feel, size, topics, length, and quotes or phrases to use, or to start with a joke. Here is a crazy prompt I put into Decktopus to create a 10-slide deck. "Create a new employee onboarding process using minimalist, sleek design with highlights of rust color that is creative not stiff, and professional." I answered four more questions about what I wanted, and this is what it gave me. decktop.us/qrBxXebhd Copy this link and paste it into your browser bar. Now, this is a stupid example. I just wanted you to see what AI slide deck creators can do with one sentence and four questions.
You can use this slide deck as your course's outline/first draft before you start adding Storyline or Rise functionality.
THEN... it could get really interesting. Imagine being on a slide, opening up an input AI Chatbot in Storyline, and just typing what you want that slide to do. The AI Chatbot writes the triggers, variables, layers, and states for you. You copy and paste it into the various places (Obviously, Articulate would need to create the places for the code to go for the various functions above), and then the slide is ready for you to check to make sure it is functioning the way you "prompted" the AI Chatbot. Obviously, the better your prompt is, the better your output is going to be for your project.
What if you could create an entire course in Storyline using a Chatbot question and answer session format? Yes, you might need to answer 200 questions with some sub-questions thrown in. So, maybe it takes 1.75 hours for that session; Storyline could write the whole course after that intake session in 20 minutes to 40 minutes. So, let's just say it took you 2.5 hours to have a completed course and 4 hours to check it and tweak it a little. That adds up to 7.5 hours. Wouldn't that be awesome?! For you workaholics that skip lunch and work 9.5 hours or more a day, it means you could go to your morning meeting to discuss a new course you want to create, do the Storyline intake creator, generate the course, check it over, debug it, tweak it for four hours or so, and it would be ready to deliver by 7:30 pm or earlier that same day. I mean, actually, in a deployable state. Articulate, this is where your industry is going. Some e-learning platform will be able to do this within the next six and a half years. Why not you?
ChatGPT will create Javascript code for you now and an entire usable APP, and you can ask it how to put it into your project if you need coding for a function and you don't want to wait on your IT department or pay to have it written. Yes, you are going to need to tweak it, but it is better than nothing.
Here is a post by Zapier all about AI platforms and what they do. https://zapier.com/blog/best-ai-productivity-tools/ I know nothing about Zapier and have zero affiliation with any AI company, so everyone knows.
Articulate, your company needs to start using AI within the company now! Whether you put it into your products now is one thing, but if your company isn't using AI to gain a competitive advantage in course creation and e-learning for your internal functions, it is probable (not just possible) you won't be around in 10 years. In November 2022, a new type of data chip was developed. This new data chip enabled hardware to store much larger amounts of data and sift through that data. It was the AI industry game changer, according to what I read.
Yes, artificial intelligence (AI) can be very useful for e-learning design and development. Here are some of the ways AI can be used in e-learning:
Personalization: AI can be used to personalize e-learning content for each learner. This can be done by tracking learner progress, interests, and learning styles, and then delivering content that is tailored to each individual learner's needs.
Adaptive learning: AI can also be used to create adaptive learning experiences. This means that the content and difficulty of the e-learning can be adjusted based on each learner's performance. This can help learners to stay engaged and challenged, and it can also help them to learn more effectively.
Recommendation engines: AI can be used to create recommendation engines that suggest relevant e-learning content to learners. This can help learners to find the content that is most relevant to their needs, and it can also help them to discover new content that they may not have found otherwise.
Assessment: AI can be used to automate assessment in e-learning. This can free up instructors to focus on other tasks, and it can also help to ensure that assessments are fair and accurate.
Chatbots: AI can be used to create chatbots that can answer learner questions and provide support. This can help learners to get the help they need when they need it, and it can also help to reduce the workload on instructors.
Overall, AI has the potential to revolutionize e-learning. By automating vehicle tracking system tasks, personalizing content, and providing personalized support, AI can make e-learning more effective, engaging, and efficient.
I mentioned this in another thread, here's an example of using generative tools (midjourney and Adobe's generative fill) to create this labeled info graphic. I modified the hot spots to make them smaller.
Indeed, ChatGPT has been a great help for saving time researching, comparing content theories, perspectives... Also for creating exercises and scenarios, just giving it context and it comes up with great ideas... No doubts it is a great and very fast extension of our skills and knowledge. It would be great to see Articulate including an AI Translation Generator for courses (in Rise, for example).
Interested in machine translation features. Powerpoint offers built in translation tools for 60+ languages. Would hate for that to take over as a de facto authoring tool.
Great topic! I've dabbled with AI tools like ChatGPT for e-learning content and it's been a game-changer for me. It helps streamline the process and adds an interactive element. Would love to hear others' experiences too!
Interested to know more about future AI capabilities in Both Storyline and Rise. My organization is tasking me to look at features and evaluate if Articulate is still the right platform for our instructional design needs. Any new word on new robo-voices, chatbots etc.
Seems like co-pilot/centaur/cyborg human + AI work is the future. The only thing that can beat a chess AI is a human and chess AI working together. eLearning will eventually have its version. Already seeing that in call center operations.
In the mean time... high quality robo voice to allow for multiple authors easily create and update courseware VOs would be super helpful. -And auto translation features would be amazing, especially if it includes the VO component.
Not holding my breath, but that's the need. Implementing work arounds in the mean time.
39 Replies
Lee-Ann, you raise an interesting point. The answer is going to depend on the license terms under which you obtained the photo.
I am not a lawyer, but my understanding is that most stock photo companies allow you to modify (non-editorial) royalty-free images they license to you, subject to certain restrictions (e.g., not making it seem like the person in the photo is supporting a particular political position, or creating "harmful" content).
In practical terms, it would seem like it makes no real difference if I edit the photo or the AI edits the photo, but I was surprised to see that at least one large stock photo company has a clause in their license agreement specifically addressing AI use.
Unfortunately, it's written just vaguely enough that I can't quite tell what it means. It *seems* (to my untrained, non-lawyer eyes) to be saying you can't use their images to TRAIN an AI. I'm not really sure if lawyers would consider using the image in a Midjourney prompt as training the AI (I wouldn't consider it to be such a use, but I'm not a lawyer). There's also the question of whether letting the AI see it would be considered redistribution of the stock photo (something that's usually not allowed).
Taking a super-strict/restrictive stance that disallows any use of stock photos with AI tools would ultimately be counterproductive to the stock photo companies, in my opinion. What would happen is that this would create an incentive to just use the AI tools without the stock photo input. Once the AI tools are capable enough that we can get the specific image we want for our use case from the AI, why would we need the stock photo company at all? But this very fear is likely to cause overreaction from the stock photo companies and an attempt to maintain the status quo by disallowing AI use in the short term.
There is also an understandable fear among photographers and artists that supply photos, illustrations, vector drawings, and other images to the stock photo companies that this source of income is under threat from AI art tools. There is a movement among artists to require AI companies to obtain the artists' permission before using any of the artists’ images to train an AI. If enough artists band together and are able to establish this as a new rule, then it will become embedded in license agreements and laws. In the meantime, stock photo companies may feel an obligation to the artists who feed images into their libraries to take stances such as “no AI use” that the artists want, even if in the long term, such a stance is counterproductive to the company.
All of which is just a long-winded way of saying that it's going to be interesting times for a while until all this settles out.
Lee-Ann - that's a really important point you make. This whole area is currently a bit of a minefield.
Legal concerns have been raised by the major stock image providers and there have already been discussions between Open AI and the image providers, as It seems likely that the 650 million or so images used for training AI text to graphics engines does include some copyright material - even though Open AI says it is taking, and has taken steps, to avoid this. But even if this issue is resolved (and it is likely to be so amicably), it still leaves the question you raise.
Personally, we are not going risk it. I think taking an image purchased as royalty-free under the terms of a usage licence does allow some latitude to modify that image before you use it in a project - I mean Shutterstock, itself, allows you to filter it, change colour, crop it etc. when you download it! But using it as input to generate an entirely new image - or set of images using AI may well be in breach of the terms of that licence.
My assertion that the dispute between image providers and Open AI will be resolved amicably is born out by the recent partnership announced by Shutterstock and DALL-E. Shutterstock will now be offering an AI generated image service! What the cost per image will be I don't know but looking at my current spend with Shutterstock, I doubt it will be cheap! But this seems the way it will go...
Re-reading my reply, I should perhaps add that most of us buy our stock images under a "single-use" licence as it is much cheaper. i.e. if you purchase images using a Shutterstock monthly "plan" and download an image under that plan you are purchasing a licence to use that image for a single project.
We are always at pains to point this out to clients when we supply source files (which is most of the time). They can modify the course and use selected images from the course as thumbnails to promote that course BUT they are in breach of copyright if they then take an image from one of our courses and use it in a different course without paying a new licence fee
To purchase an image "royalty-free" is generally way more costly and we do it only when the client requires that image for multiple projects.
I have recently been very interested in AI. Five days ago, I really started to delve into the topic and think about how it could help with Storyline. All the ideas listed here are great, but AI can do much more with course creation. For example, you can currently use Decktopus, Beautiful.ai, or Slidesgo to create a custom slide deck of your topic all by typing into a box your parameters like colors, look and feel, size, topics, length, and quotes or phrases to use, or to start with a joke. Here is a crazy prompt I put into Decktopus to create a 10-slide deck. "Create a new employee onboarding process using minimalist, sleek design with highlights of rust color that is creative not stiff, and professional." I answered four more questions about what I wanted, and this is what it gave me. decktop.us/qrBxXebhd Copy this link and paste it into your browser bar. Now, this is a stupid example. I just wanted you to see what AI slide deck creators can do with one sentence and four questions.
You can use this slide deck as your course's outline/first draft before you start adding Storyline or Rise functionality.
THEN... it could get really interesting. Imagine being on a slide, opening up an input AI Chatbot in Storyline, and just typing what you want that slide to do. The AI Chatbot writes the triggers, variables, layers, and states for you. You copy and paste it into the various places (Obviously, Articulate would need to create the places for the code to go for the various functions above), and then the slide is ready for you to check to make sure it is functioning the way you "prompted" the AI Chatbot. Obviously, the better your prompt is, the better your output is going to be for your project.
What if you could create an entire course in Storyline using a Chatbot question and answer session format? Yes, you might need to answer 200 questions with some sub-questions thrown in. So, maybe it takes 1.75 hours for that session; Storyline could write the whole course after that intake session in 20 minutes to 40 minutes. So, let's just say it took you 2.5 hours to have a completed course and 4 hours to check it and tweak it a little. That adds up to 7.5 hours. Wouldn't that be awesome?! For you workaholics that skip lunch and work 9.5 hours or more a day, it means you could go to your morning meeting to discuss a new course you want to create, do the Storyline intake creator, generate the course, check it over, debug it, tweak it for four hours or so, and it would be ready to deliver by 7:30 pm or earlier that same day. I mean, actually, in a deployable state. Articulate, this is where your industry is going. Some e-learning platform will be able to do this within the next six and a half years. Why not you?
ChatGPT will create Javascript code for you now and an entire usable APP, and you can ask it how to put it into your project if you need coding for a function and you don't want to wait on your IT department or pay to have it written. Yes, you are going to need to tweak it, but it is better than nothing.
Here is a post by Zapier all about AI platforms and what they do. https://zapier.com/blog/best-ai-productivity-tools/ I know nothing about Zapier and have zero affiliation with any AI company, so everyone knows.
Articulate, your company needs to start using AI within the company now! Whether you put it into your products now is one thing, but if your company isn't using AI to gain a competitive advantage in course creation and e-learning for your internal functions, it is probable (not just possible) you won't be around in 10 years. In November 2022, a new type of data chip was developed. This new data chip enabled hardware to store much larger amounts of data and sift through that data. It was the AI industry game changer, according to what I read.
This post was removed by the author
Yes, artificial intelligence (AI) can be very useful for e-learning design and development. Here are some of the ways AI can be used in e-learning:
Overall, AI has the potential to revolutionize e-learning. By automating vehicle tracking system tasks, personalizing content, and providing personalized support, AI can make e-learning more effective, engaging, and efficient.
Great points Lucia and David, the lists seems to be a good feature request entry.
Good AI responses.
@sarah hodge, would love to be updated from time to time if and when Articulate starts adding these in the roadmap, even if little by little.
Almost all platforms are in the AI arms race now. Curious to see Articulate's next move.
Thanks all.
I mentioned this in another thread, here's an example of using generative tools (midjourney and Adobe's generative fill) to create this labeled info graphic. I modified the hot spots to make them smaller.
Indeed, ChatGPT has been a great help for saving time researching, comparing content theories, perspectives... Also for creating exercises and scenarios, just giving it context and it comes up with great ideas... No doubts it is a great and very fast extension of our skills and knowledge. It would be great to see Articulate including an AI Translation Generator for courses (in Rise, for example).
Great example of how to use ChatGPT!
Interested in machine translation features. Powerpoint offers built in translation tools for 60+ languages. Would hate for that to take over as a de facto authoring tool.
Hey Ashley,
Great topic! I've dabbled with AI tools like ChatGPT for e-learning content and it's been a game-changer for me. It helps streamline the process and adds an interactive element. Would love to hear others' experiences too!
Cheers,
Sandra
can you share?
Interested to know more about future AI capabilities in Both Storyline and Rise. My organization is tasking me to look at features and evaluate if Articulate is still the right platform for our instructional design needs. Any new word on new robo-voices, chatbots etc.
Seems like co-pilot/centaur/cyborg human + AI work is the future. The only thing that can beat a chess AI is a human and chess AI working together. eLearning will eventually have its version. Already seeing that in call center operations.
In the mean time... high quality robo voice to allow for multiple authors easily create and update courseware VOs would be super helpful. -And auto translation features would be amazing, especially if it includes the VO component.
Not holding my breath, but that's the need. Implementing work arounds in the mean time.