Priya Rao
Priya Rao @priya-r · 2 days ago
Discussion

Midjourney V5.2 Copyright: A Misguided Debate

Unpopular opinion: The current copyright debates surrounding AI art generated by Midjourney v5.2, particularly regarding the use of copyrighted training data, fundamentally misunderstand the creative process and demand a more nuanced legal framework.
▲ 6 upvotes 💬 3 replies ← Back to Community

3 Replies

Aisha R.
Aisha R. @aisha-r · 2 days ago ▲ 2
I totally agree – it feels like the focus on training data ignores how users actively shape the output with prompts like Midjourney’s “stylize” parameter, which I’ve seen drastically change images with just a setting of 7.
Lisa M.
Lisa M. @lisa-m · 2 days ago ▲ 2
The core issue isn’t just training data; Midjourney’s ‘Remix’ feature allows for 10-20 variations based on prompts, creating a significant gray area we haven’t addressed legally. We need to focus on prompt provenance and demonstrable influence, not simply blanket restrictions on dataset usage.
Marcus Davis
Marcus Davis @marcus-d · 3h ago ▲ 2
I've been experimenting with Midjourney V5.2’s “Variations” feature – generating 100 iterations based on a single prompt yielded 37 distinct images with demonstrable stylistic overlaps resembling specific artists’ work, suggesting a significant risk of infringement.
Join the discussion

Sign in to reply, vote, and connect with the AIZyla community.

Join Community →

Related discussions