Responsible AI

Unelma builds creative tools powered by AI. This is how we build, deploy and govern them responsibly.

Our principles

• Human-centered creativity — AI augments imagination; it doesn't replace the designer's authorship or intent.
• Transparency — we're clear when content has been generated or assisted by AI.
• Fairness — we work to reduce bias and treat diverse aesthetics, bodies and cultures with respect.
• Safety — we prohibit content that exploits, defames or infringes the rights of others.

Training data

Our models are trained on curated datasets. We respect intellectual property and make reasonable efforts to avoid using copyrighted material without authorization. You keep full ownership of what you generate.

What we prohibit

• Generating content that impersonates a real person without consent.
• Producing content that's unlawful, hateful, sexualizes minors, or otherwise harmful.
• Copying or closely replicating a third party's protected designs or trademarks.

Human oversight

Every design starts from a human prompt and stays under your control. You can edit, refine or discard any AI output before publishing.

Report a concern

If you encounter harmful, biased or infringing content generated on Unelma, write to support@unelma.dev. We review reports and may remove content or restrict accounts that break these principles.