Forum Discussion
Office Hours #2 Discussion
Category #2: AI Images
If you do get an image with extra hands or fingers or things you don't want, what can I add to the prompt to fix these types of issues?
Is there a way to save the custom prompts you have created within Rise or Storyline?
Is there a way to upload an image as reference for the generation of the new image?
When i asked in the last office hour, I was told no. it would be so great to do this.
This would be a great feature to have.
GQuicano​ LeeMillardButlr​ MoniqueGuggi738​ something like this is coming to Storyline. Check out what David shared about AI Image Editing (an experimental feature) at the 43-minute mark (43:00) in this webinar recording.
THANK YOU AlyssaGomez​ ! I was literally trying to find the recording of this webinar today. I thought I remembered seeing this during this webinar.
I normally use Nano Banana Pro for consistent scenario character image creation. How does the built-in AI image creation compare in the same situation (same character added to different contexts, facial expressions, outfits, adding additional characters, etc.)?
LeeMillardButlr​ The AI Assistant is on it's way there. Check out what David shared about AI Image Editing (an experimental feature) at the 43-minute mark (43:00) in this webinar recording.
What are some effective ways to avoid common issues in AI-generated images, like people having multiple faces, extra fingers, or duplicated individuals who all look the same (e.g., identical clones around a meeting table)?
How can you write prompts to minimize these problems? Are there specific techniques or wording that help improve accuracy and consistency?
Any tips would be really appreciated. I’d love to avoid wasting tokens on unusable results.
This is a great question NadiaChernyak1​ and I'm curious to hear what techniques others in this cohort would recommend.
When it comes to hallucinations (extra fingers, multiple faces), my best recommendation is to just regenerate new images without changing the prompt. Also, be sure you're using the latest image generator model, since you'll the get best results with that.
Are there any copyright or licensing concerns when using AI-generated images in courses distributed to clients or employees?
AnnaSantiago​ You can find answers to copyright questions in our Legal and Compliance FAQs.
When I asked for a diverse group of people, it would only give me black and white. How do I get it to give me more diversity without having to spell things out. I'd love so see people with disabilities, religious, cultural, tattoos, colored hair and so on.
MoniqueGuggi738​ my recommendation is to start more general in your prompt, then get more specific each time you generate.
Though I agree, diversity shouldn't be something that we should ask for. It's something that should be baked into the system.
I loved the different custom AI photo styles that were shown in the training! Is there a resource you could point me to so I can find some prompt ideas for creating my own prompt library?
Yes, check out this video ElizabethArchib​!
I beleive someone may have spoke on this or it may have been in the training, but is there a way to generate a series of related pictures you can use throughout a course.
Based on last office hour, no :-(
ShimellH​ LeeMillardButlr​ MoniqueGuggi738​ you can view what David shared about AI Image Editing (an experimental feature) at the 43-minute mark (43:00) in this webinar recording.
I think this feature is forthcoming, based on the recent "What's New in Articulate 360" webinar.
