Legal

AI at InksForm

Version 1.3, effective 10 October 2026 · History

In short

  • InksForm uses AI to create images, video, voice, text and websites. Assistants always say they're AI.
  • Pictures and videos made with Google carry Google's invisible SynthID watermark, and every file made with AI in InksForm says so in its own data.
  • We don't train AI models on your content.
  • When you publish realistic AI content about real people or events, label it.

The AI we use

  • Text, website code, the assistant and research: Claude by Anthropic.
  • Images and video: Google Gemini and Veo (Gemini also reads the audio and video you add).
  • Voice, music, sound and transcription: ElevenLabs.
  • Your own Claude key (optional): the website Builder and the assistant can use your own Anthropic account instead of credits.

Each provider processes your prompt and chosen files only to produce your result. See the Privacy policy.

How results are marked

  • Images and videos made with Google Gemini and Veo carry Google's invisible SynthID watermark, and Google's Content Credentials (C2PA) in the original file.
  • Edits, designs, memes and video exports are redrawn in your browser, which removes Google's Content Credentials. These files carry our IPTC mark instead: it says the file was made or edited with AI and names the AI, never your prompt or account. Signed Content Credentials of our own are coming.
  • PDF, Word and PowerPoint files with AI-written text say so in their document properties.
  • Voice and music don't carry a machine-readable mark yet.
  • Anyone can check a file: it's read on their device, not uploaded.
  • Chat assistants in InksForm always make clear they're AI.

This follows Article 50 of the EU AI Act, which has applied since 2 August 2026.

Your part when you publish

  • Label realistic AI images, audio or video of real people, places or events ("deepfakes") as AI-generated.
  • If you publish AI-written text to inform the public on matters of public interest, say it was AI-generated unless a person reviewed and took editorial responsibility for it.
  • Don't remove or alter AI marks on files.

Limits

AI can be wrong, biased or out of date. Check results before relying on them, especially facts, prices, names and anything legal, medical or financial. You stay responsible for what you publish.