AI Content Policy

Last updated: [TO BE SET]

1. Transparency

ABMediaX Poster (the "Service") generates content using artificial intelligence. Depending on which feature you use, the Service may generate videos, audio voiceovers, captions, hashtags, and post copy. When you publish through the Service, some or all of the media may be synthetic or materially altered by AI. This page describes how we handle that.

2. Platform Labeling — Automatic

Each major target platform now requires that AI-generated or materially altered media be labeled. When the source media for a published post was created by the Service (i.e. the underlying media_asset has source = 'generated'), ABMediaX Poster automatically sets the platform-specific AI-content flag on the outgoing API call, so you are compliant by default:

  • TikTok: the Direct Post API is called with post_info.disclose_ai_generated_content = true (and disclose_commercial_content = true when a workspace member has flagged the post as an ad).
  • Meta (Instagram Reels): the Content Publishing API media-create call is issued with ai_labeled = true, which surfaces Meta's"Made with AI" disclosure on the resulting Reel.
  • YouTube: the videos.insert request sets the "Altered content" disclosure (aiGeneratedContent = true on the video resource), which YouTube surfaces to viewers of Shorts.
  • X: X currently exposes no first-party API flag for AI-content disclosure. When required by X policy, you must include a textual disclosure in the post itself.

You are ultimately responsible for confirming that every post you publish complies with the disclosure rules of the platform on which it appears. If a platform's policy changes we will update the adapters as soon as reasonably practicable.

3. Watermarking & Provenance

Every AI-generated video output from the Service is stamped with a C2PA-style provenance metadata tag that identifies the content as AI-generated and records the model and timestamp used. You may not remove or alter this provenance data. Attempting to strip provenance is a violation of the Acceptable Use Policy.

4. Copyright of AI Outputs

Under current U.S. Copyright Office guidance (Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16190 (Mar. 16, 2023)), works produced solely by generative AI without meaningful human creative contribution are generally not copyrightable. Works with substantial human authorship — for example, a video where you scripted, directed, and curated the AI outputs — may be copyrightable as to that human contribution.

You assume the risk that some or all of your AI Outputs may be uncopyrightable. As between you and ABMediaX, we assign to you all right, title, and interest we may have in the AI Outputs, to the extent transferable under applicable law (see Terms § 6.1).

5. Prohibited AI Uses

The following are prohibited (also see AUP § 2):

  • Deepfake video, image, or voice of a real person without prior written consent, logged in your workspace via /dashboard/consent;
  • Election-related synthetic media without a clear and conspicuous disclosure that the content is AI-generated;
  • Non-consensual sexual or intimate imagery of any person;
  • Impersonation of public figures, government officials, or brand personas outside of obvious satire or commentary in the public interest;
  • Misleading advertising, fake reviews, fabricated testimonials, or fabricated news;
  • Attempts to defeat safety filters, provenance tags, or content-classifier flags.

6. Real-Person Consent

Before the Service will plan or render a video prompt that names a specific real person, the workspace must have on file a completed real-person consent record for that individual. You submit consent at /dashboard/consent. Records include:

  • The subject's legal name;
  • An uploaded consent document (PDF or image) — for example, a signed release;
  • An attestation you type yourself confirming that the consent is current and specific to AI-generated depictions;
  • Your IP address and user agent at time of submission.

Consent records are retained for 5 years after the last related post, so that we (and you) can respond to complaints or subpoenas. See our Privacy Policy for details on retention.

7. Model Training

We do not use your prompts, uploads, or generated outputs to train AI models. Our AI subprocessors — OpenAI, Runway, ElevenLabs — process your inputs on their commercial APIs, where their standard terms exclude API traffic from model training.

8. Contact

Concerns about AI content generated by the Service: legal@abmediax.com. Reports of prohibited AI use by others: abuse@abmediax.com.