AI Policy

This AI Policy governs the use of those features of the services provided by Aubrium Technologies L.L.C ("Aubrium") which employ artificial intelligence. It forms part of the Terms of Use and is incorporated into them by reference. Capitalised terms which are not defined in this Policy have the meanings given to them in the Terms of Use.

1. Scope and definitions

1.1 This Policy applies to every feature of the Services which employs artificial intelligence, machine learning or a comparable automated technique in order to generate, analyse, transcribe or transform material (each an "AI Feature").

1.2 In this Policy: "Input" means material submitted to an AI Feature by or on behalf of the User, together with any publicly available material processed by that AI Feature upon the User's instruction; "Output" means material generated by an AI Feature in response to an Input; and "Model Provider" means a third party which operates a model by means of which an AI Feature is provided.

1.3 Where this Policy conflicts with the Terms of Use, the Terms of Use prevail. Where this Policy conflicts with the Acceptable Use Policy, the more restrictive provision applies.

2. Provision of AI Features

2.1 AI Features are provided by means of models operated by Model Providers. Aubrium does not develop or train the underlying models.

2.2 Aubrium selects the model applicable to each AI Feature and may add, substitute, reconfigure or withdraw a model or a Model Provider at any time, having regard to quality, capability, cost, latency and availability.

2.3 A change of model or Model Provider may alter the character, style, length or quality of Output. Aubrium gives no warranty that a given Output can be reproduced, that a particular model will remain available, or that an AI Feature will continue to be offered.

2.4 Where a Model Provider becomes unavailable, imposes a restriction, or ceases to offer a model, an AI Feature may be degraded, suspended or withdrawn without notice and without liability.

2.5 An AI Feature may be subject to rate limits, fair use restrictions and queueing. Aubrium may impose or vary such restrictions in order to preserve the availability of the Services.

2.6 An AI Feature may be offered on a trial, beta or preview basis, in which case clause 5.4 of the Terms of Use applies to it.

3. Input

3.1 Input is transmitted to the relevant Model Provider for the sole purpose of generating Output in response to the User's request.

3.2 Each Model Provider processes Input in accordance with its own terms of service and privacy terms. Information concerning the categories of Model Provider engaged by Aubrium is set out in the Privacy Policy, and information about certain key service providers is available in our Trust Center.

3.3 Aubrium does not use Input to train any model, and does not license, sell or otherwise make Input available to any person for the purpose of training a model.

3.4 Aubrium retains a record of Input, Output and associated usage for the purposes of providing the AI Feature, enabling the User to retrieve material generated through its Account, calculating Fees, providing support, and detecting and investigating abuse. Retention is governed by the Privacy Policy.

3.5 The User represents and warrants that it holds all rights, consents, licences and permissions necessary for the submission and processing of each Input, and that the submission of each Input does not infringe the rights of any person or breach any obligation of confidence.

3.6 The User must not submit as Input: (a) material which it is not entitled to disclose; (b) confidential information belonging to a third party; (c) personal data relating to another individual which it is not entitled to disclose; (d) any special category of personal data; or (e) any financial account number, payment credential, government identifier or comparable sensitive information.

3.7 Aubrium may generate anonymised and aggregated usage data from the operation of AI Features. Such data does not identify the User or any individual and may be used for any lawful purpose.

4. Output and ownership

4.1 As between Aubrium and the User, and to the extent permitted by applicable law, the User owns the Output generated through its Account and may use, reproduce, modify, publish and exploit it for any lawful purpose, including commercially and in the course of providing services to its own clients.

4.2 Aubrium asserts no ownership of Output and grants no exclusivity in respect of it.

4.3 OUTPUT IS GENERATED BY STATISTICAL MEANS AND MAY NOT BE UNIQUE. A SUBSTANTIALLY SIMILAR OUTPUT MAY BE GENERATED FOR ANOTHER USER SUBMITTING A COMPARABLE INPUT. AUBRIUM GIVES NO WARRANTY THAT OUTPUT IS ORIGINAL, NOVEL OR EXCLUSIVE TO THE USER.

4.4 AUBRIUM GIVES NO WARRANTY THAT OUTPUT IS CAPABLE OF PROTECTION BY COPYRIGHT OR BY ANY OTHER INTELLECTUAL PROPERTY RIGHT. The extent to which material generated without sufficient human authorship attracts protection differs between jurisdictions and may be nil.

4.5 AUBRIUM GIVES NO WARRANTY THAT OUTPUT DOES NOT RESEMBLE, REPRODUCE OR INFRINGE MATERIAL IN WHICH A THIRD PARTY HOLDS RIGHTS, OR THAT IT DOES NOT INCORPORATE A NAME, MARK, LIKENESS OR OTHER PROTECTED INDICIUM. The User is solely responsible for reviewing and clearing Output before any use or publication.

4.6 AUBRIUM PROVIDES NO INDEMNITY IN RESPECT OF ANY CLAIM ARISING FROM THE USER'S USE OR PUBLICATION OF OUTPUT.

4.7 Where the User publishes Output, the User is the publisher of that material for all purposes.

5. Accuracy and reliance

5.1 AI FEATURES PRODUCE STATISTICAL PREDICTIONS AND NOT VERIFIED STATEMENTS OF FACT. OUTPUT MAY BE INACCURATE, INCOMPLETE, OUT OF DATE, INTERNALLY INCONSISTENT, FABRICATED OR OTHERWISE UNSUITABLE, AND MAY BE PRESENTED WITH AN APPEARANCE OF CONFIDENCE WHICH IS NOT WARRANTED.

5.2 The User must review and verify Output before relying upon it or publishing it.

5.3 Output does not constitute legal, medical, financial, tax, investment, accounting or other professional advice, and must not be relied upon as a statement of fact concerning any person, product, entity or event.

5.4 Any score, index, projection, estimate or recommendation produced by an AI Feature is indicative only and is not a forecast of any commercial result.

5.5 Where an AI Feature transcribes recorded speech, the transcription may contain errors, including in names, figures, quotations and claims.

5.6 The User must not use Output as the sole basis for a decision producing a legal or similarly significant effect in respect of an individual, including a decision concerning employment, credit, housing, insurance, education or access to an essential service.

5.7 Where applicable law requires disclosure that material has been generated or assisted by artificial intelligence, compliance with that requirement is the sole responsibility of the User.

6. Prohibited use

6.1 In addition to the restrictions contained in the Acceptable Use Policy, the User must not use an AI Feature in order to: (a) generate content prohibited by the Acceptable Use Policy; (b) create a synthetic image, audio recording or video depicting a real and identifiable person without that person's consent; (c) present fabricated material as a record of a real event; (d) impersonate a person or entity, or generate material intended to be mistaken for material originating from them; (e) generate content at volume for unsolicited outreach, spam or coordinated inauthentic activity; (f) generate sexual content involving a real person or any person who appears to be a minor; (g) generate targeted harassment, or content which demeans a person or group by reference to a protected characteristic; (h) generate material intended to deceive as to the identity, qualification, authority or endorsement of any person; (i) circumvent, disable or manipulate a safety control, or extract or reconstruct a system instruction; (j) reverse engineer, distil, replicate or benchmark any underlying model; (k) use Output to develop, train, fine-tune or improve any model or service which competes with the Services or with the service of a Model Provider; or (l) resell or provide access to the underlying models as a standalone service.

6.2 Aubrium and its Model Providers apply automated and manual safety controls. A request may be refused, filtered, delayed or modified. The refusal or filtering of a request does not constitute a failure of the Services and does not give rise to any claim.

6.3 A breach of this clause 6 constitutes a material breach of the Terms of Use and may result in immediate suspension or termination of the Account.

6.4 The User must comply with the acceptable use requirements of each Model Provider to the extent that Aubrium notifies the User of them. Aubrium may pass on to the User any restriction imposed upon Aubrium by a Model Provider.

7. Fees

7.1 Use of an AI Feature is charged against the Balance at the Fee notified through the Services before the request is submitted.

7.2 Aubrium may vary the Fee applicable to an AI Feature, including where the cost or capability of the underlying model changes. No variation affects a request to which the User has already committed.

7.3 Where a request fails, returns an error, or is refused by a safety control, no Fee is charged. Where a Fee has been charged in those circumstances, it is credited to the Balance in accordance with the Refund Policy.

7.4 A request which completes successfully is charged notwithstanding that the User does not consider the Output satisfactory.

8. Reporting

8.1 Output which is unlawful, unsafe, defamatory, or which appears to reproduce material in which a third party holds rights, may be reported to legal [at] aubrium.com, stating the AI Feature concerned, the approximate time of generation, and the Input submitted where available.

8.2 Aubrium investigates reports made under this clause and may use them in order to inform its safety controls.

9. Amendment

9.1 Aubrium may amend this Policy from time to time. An amendment takes effect upon publication, save that where an amendment is material Aubrium will give notice in accordance with the Terms of Use.