Definitions
- Input means the prompts, brand settings, instructions, references and files you give us.
- Output means the text, images, video, audio and other material the Service generates from Input.
Who owns the Output
As between you and Axis, and provided your account is in good standing, you own the Output you generate. To the extent we hold any right in it, we assign that right to you.
Limitation on protectability. Material generated by artificial intelligence may not attract copyright protection in the United States or in other jurisdictions, as protection generally requires human authorship. You may therefore be unable to prevent third parties from using materially similar material. The degree of human direction, selection and editing applied may affect that position. Axis makes no representation as to the protectability of any Output.
Output may not be unique
The underlying models are shared across all customers. Another customer submitting comparable instructions may receive similar or materially identical results. You acknowledge and accept this, and no claim arises between customers, or against Axis, on that basis. Axis gives no undertaking of exclusivity in respect of any Output.
Output may be wrong
AI systems produce confident text that can be inaccurate, out of date, internally inconsistent or entirely invented. They can also produce material that resembles existing work.
Axis does not warrant that Output is accurate, current, complete, original, non-infringing or fit for any purpose. Output does not constitute professional advice of any kind.
You review before you publish
Output must be reviewed by a natural person before it is published or relied upon. That responsibility rests with you. Scheduling material through the Service does not transfer it to Axis.
You are responsible for checking that Output is accurate, that any claim in it can be substantiated, that it complies with the law and with the rules of the channel it is going to, and that it does not infringe anyone's rights.
Labeling and provenance
Two distinct obligations apply, and they rest with different parties.
| What it is | Who controls it | |
|---|---|---|
| Embedded provenance | Machine-readable metadata inside every image and video you download or publish, recording that it was AI-generated and by which model. It is invisible, it does not appear on the image, and it does not change the caption. | Axis. Always applied. Not a setting. The duty to mark synthetic media sits with the provider of the AI system, which is us. |
| Visible disclosure | A short line published alongside your post saying the content was AI-assisted. | You. A setting you turn on or off per channel. We default it on for accounts in the European Union and the United Kingdom, and off elsewhere. You can change it at any time. |
You are responsible for determining whether a visible disclosure is required in respect of your audience, your industry and your jurisdiction, and for enabling it where it is. Disabling the setting does not remove the embedded provenance data and does not transfer any disclosure obligation to Axis.
What we train on
Axis may use for training: your prompts, brand settings and instructions; Output generated for you; and the signals arising from your use of that Output, including which drafts are retained, edited, discarded or published.
Axis does not use for training: any material retrieved from a connected platform, including followers, audience data, comments, direct messages and performance metrics; payment and billing information; or the contents of support correspondence.
The first exclusion is also a contractual requirement. Meta, TikTok and Google each prohibit the use of data obtained through their platforms for the training of artificial intelligence models, and Axis complies with those restrictions.
Withdrawal may be effected in your account settings or by written notice to privacy@axismarketing.ai. Withdrawal takes effect prospectively. Content already incorporated into a trained model cannot be withdrawn from that model, but Axis will exclude your content from all subsequent training.
Model providers and change
We use AI models supplied by third parties and we choose which model handles which task. We may change models, providers or configuration at any time to improve quality, cost or reliability. Output for the same Input may therefore differ over time.
Providers impose their own usage rules, which reach you through the Acceptable Use Policy. Where a provider restricts something, we must restrict it too.
Prohibited uses
The Acceptable Use Policy governs what may be generated. Two limits matter most here:
- No real people. Axis does not generate images, video or audio of real, identifiable people, and does not accept a real person's likeness or voice as a reference.
- No third-party brands. No other company's trademarks, logos or trade dress, and no naming real brands in your instructions.