I wanted to share the first draft of Signum, an open standard I've been putting together for labeling AI involvement in digital content.
I decided to put this standard together as I found myself wanting to display how much I'd leveraged AI when writing code, documentation, or creating digital media content.
Looking around, I couldn't find existing way to add this information to my content that was consistent and standardised, enter Signum.
What Signum Is:
Signum proposes a simple, 5-level framework (from purely Human to Fully Automated AI) with associated metadata and optional visual cues to label content based on the creation process.
The goal isn't to judge quality, but simply to provide transparency about how something was made.
Seeking Feedback & Guidance:
This is genuinely my first time attempting to build out and share a standard/project like this.
I'm fully aware it's likely naive in places and definitely incomplete.
I'm putting this draft out specifically because I'm keen to get constructive feedback, criticism, suggestions for improvement, pointers to things I've missed, and guidance from the HN community.
If you think the approach is flawed, or could be better, I'd love to hear why. Collaboration would be fantastic too!
I wanted to share the first draft of Signum, an open standard I've been putting together for labeling AI involvement in digital content.
I decided to put this standard together as I found myself wanting to display how much I'd leveraged AI when writing code, documentation, or creating digital media content.
Looking around, I couldn't find existing way to add this information to my content that was consistent and standardised, enter Signum.
What Signum Is:
Signum proposes a simple, 5-level framework (from purely Human to Fully Automated AI) with associated metadata and optional visual cues to label content based on the creation process.
The goal isn't to judge quality, but simply to provide transparency about how something was made.
Seeking Feedback & Guidance:
This is genuinely my first time attempting to build out and share a standard/project like this.
I'm fully aware it's likely naive in places and definitely incomplete.
I'm putting this draft out specifically because I'm keen to get constructive feedback, criticism, suggestions for improvement, pointers to things I've missed, and guidance from the HN community.
If you think the approach is flawed, or could be better, I'd love to hear why. Collaboration would be fantastic too!
Appreciate you taking the time to look.
Thanks,
CT