AI Art Assistant


I am currently looking for collaborators for a machine learning and art project. Would you like to join me? Continue reading below!

Hello future collaborator,

I am a software artist and developer looking for potential collaborators who would be interested in testing and/or using a software art tool I am designing. I am applying for grants to the Canada Council for the Arts. The software will be open source and free to use by anyone, but I’m hoping to have some support by artists and arts groups to support my grant applications.

There is no financial requirement or even strict obligation to use this tool. Instead, I would simply require a statement of interest from potential collaborators: a brief description of their practice or previous work and why they are interested in using the software. Collaborators would also be welcome to be further involved in the design or testing of the tool. If you wish to commit to helping develop the tool or creating with it, please contact me. I’m happy to answer any questions you have through email, phone or videochat.

Cheers, Ryan

Ansanm: AI Art Assistant Software:

Download a PDF detailed description of the project

Ansanm (Haitian creole for “together”, from the French “ensemble”) is ethically designed, open-source software that increases accessibility of visual storytelling and provides simple, thoughtful interaction with an artistic assistant AI. The Ansanm AI will help artists quickly construct an initial concept based on a story, text or poem which is then available for further refinement using smart search and comparison tools. The goal is to make the creation of visual media accessible to people of all skill levels and allow for diverse storytelling. For example, given a story, a visual essay with correct timing and smooth transitions will be constructed automatically from the artist’s image library, internet content and AI generated content. Images can then be replaced, removed or inserted as needed. Use cases include:

  • Music videos, video essays and video installations
  • Photo slideshows with background music
  • Poetry with integrated visuals and other talks and presentations

Ansanm’s AI assistant can translate one medium to another, the simplest being text to image. Artists can then explore new mediums, opening up avenues that were not possible before because of a lack of training, interest, or physical ability.

Ansanm will help artists explore a new form of art making that centers around human-machine collaboration and curation. The AI assistant will find or create many options for the artist-as-curator to choose from, who can then guide the AI to modify that selection until it finally fits their artistic vision.

There are some existing for-profit businesses using some basic AI for presentations and video essays but they are strongly biased towards business and monetization on social media platforms. None are open source, most are not freely available, and all have unknown privacy concerns. They do provide an example that the technology exists and illustrate the basic features and user experience:

Phase 1 use cases:

  • Photo slideshows with background music
  • Poetry with integrated visuals and other talks and presentations
  • Music videos, video essays and video installations

Phase 1 features:

  • Import and display of text, images, videos and music
  • Timeline and editing of content sequences
  • Assisted image, video and music search:
    • Search multiple copyright-friendly (creative commons, etc) databases at once
      • Assisted rights and credits, i.e. automatic credits generation
    • Content processed by AI for further tagging and categorization
    • Automated content searches given text or lyrics as input
  • Automated sequence/timeline visual content generation from text content
  • Export to video

Phase 2 use cases:

  • Video/slideshow synchronized accompaniment (to speech and/or music)
  • Live interactive VJing
  • Interactive visual novels and storytelling games

Phase 2 feature options:

  • Improved smart searches
    • Marking images as good/poor fits and re-adjusting search
    • Searching for similar images
    • De-duplication of content (e.g. hide alternate versions or copies of images)
  • Interactivity (support for keyboard, mouse, game controllers, webcam input, etc)
  • Branching sequencing (allows of different flow based on input, etc)
  • Timing controls through audio input (i.e. advance to next image/video based on audio input)
  • Dialogue systems for visual novels
    • Import of dialogue from narrative systems like Twine, Yarn, etc
  • Collaboration and sharing
    • Easy publishing of your source code and images and access to others
    • Sharing of AI models
  • Image, video and music generation