AI Use and Transparency Policy

Lira Learning LLC | Lira Studio

Effective date: July 25, 2026

Lira Learning LLC (“Lira,” “we,” “us,” or “our”) operates Lira Studio, an AI-assisted learning design platform. This AI Use and Transparency Policy explains how artificial intelligence is used in Lira Studio, how user content is handled when AI features are used, and what users should understand before relying on AI-assisted outputs.

This Policy should be read together with Lira’s Privacy Policy and Terms and Conditions.

1. How Lira Studio Uses AI

Lira Studio uses artificial intelligence to help users plan, structure, generate, revise, translate, organize, analyze, and review learning materials.

Depending on the feature selected, AI may help users:

Develop learning objectives, activities, assessments, and instructional content.
Organize information across projects and project components.
Revise or transform existing material.
Translate or adapt content.
Review content for alignment, quality, consistency, cohesiveness, rigor, readiness, or other instructional considerations.
Identify possible gaps, conflicts, or areas that may need further review.
Apply instructional context and guidance selected or provided by the user.

Lira identifies AI-assisted features within the platform. Users decide when to use those features and whether to accept, revise, reject, or replace the resulting content.

2. Human Decisions Lead the Work

Lira Studio is designed to support human learning design, not replace it.

Users make the decisions that shape a project, including decisions about:

Learners and learning needs.
Goals and intended outcomes.
Project structure and sequence.
Source material and subject matter.
Instructional methods and supports.
Standards, requirements, or other forms of alignment.
The suitability of content for a particular setting.
Final approval, implementation, and publication.

Lira may use research-informed guidance, structured project context, user instructions, and artificial intelligence to develop or review content. These tools support professional judgment. They do not replace it.

Users are responsible for reviewing and approving content before it is taught, published, distributed, submitted to a client, or used to make decisions about a learner.

3. How AI Requests Are Processed

When a user selects an AI-assisted feature, Lira may send the information needed to complete the request to an external AI service provider.

This information may include:

The user’s instructions or prompt.
Relevant project content.
Uploaded or entered source material.
Selected learner, delivery, population, safety, or readiness information.
Relevant project structure and component information.
Lira’s instructional guidance and system instructions.

Lira sends only the information reasonably needed to perform the requested task. Additional information about how personal information and project content are processed is available in our Privacy Policy.

When a user chooses an AI-assisted feature, Lira may send the instructions, relevant project content, and related context needed to complete the request to Anthropic through its commercial API. Anthropic states that commercial API inputs and outputs are not used to train its models by default and are generally deleted from its backend within 30 days, subject to product-specific retention, agreed terms, Usage Policy enforcement, feedback, and legal obligations.

4. AI Training and Use of Project Content

Lira does not use private User Content to train or fine-tune general-purpose AI models. Lira will not use identifiable private User Content to train or fine-tune any AI model without the user’s express permission or a separate written agreement. Lira may use aggregated or de-identified operational information to operate, evaluate, secure, and improve the Services.

We may also use feedback that a user voluntarily submits to us. Anthropic’s practices and Lira’s service providers may change. We will update this Policy when a change materially affects how User Content is processed. Additional details are available in Anthropic’s commercial data-retention information and model-training information.

5. Student and Sensitive Information

Lira Studio is intended for adult professionals. It is not designed for students or children to use directly.

Do not submit personally identifiable information about individual learners, including students or minors, in project content, prompts, uploads, source materials, or AI-assisted features unless Lira has expressly authorized that use in writing and appropriate legal and contractual safeguards are in place.

This includes:

Student names or initials when they could identify a student.
Personal email addresses or phone numbers.
Student identification numbers.
Education records.
Health, disability, or accommodation records.
Photographs or recordings of identifiable students.
Precise location information.
Other information that could reasonably identify an individual learner.

Users should use anonymized, aggregated, or fictional examples when learner information is needed.

Users must also avoid submitting passwords, financial account information, government identification numbers, protected health information, confidential personnel records, trade secrets they are not authorized to disclose, or other highly sensitive information.

An organization that wants to use Lira Studio with regulated or identifiable learner data must first receive written authorization from Lira and enter into any required data protection or service agreements.

6. Accuracy and Limitations of AI Outputs

Artificial intelligence can produce content that is incomplete, inaccurate, inconsistent, outdated, biased, unsupported, or unsuitable for a specific learner, purpose, or environment.

AI-generated content may:

Misstate facts or invent information.
Misinterpret source material.
Produce inaccurate citations or references.
Miss important context.
Oversimplify complex topics.
Include cultural, linguistic, demographic, or other forms of bias.
Produce content that is not age-appropriate or instructionally appropriate.
Fail to meet accessibility, legal, contractual, standards, or policy requirements.
Suggest activities that are not feasible or safe in a particular setting.

Lira’s research-informed structures and guidance are intended to improve the quality and relevance of AI-assisted work. They do not guarantee that an output is accurate, complete, unbiased, compliant, instructionally sound, or ready to implement.

Users must independently verify important facts, sources, citations, translations, standards alignments, safety considerations, and legal or contractual requirements.

7. Reviews, Evaluations, and Scores

Lira may provide reviews, evaluations, findings, ratings, scores, flags, recommendations, or other forms of instructional analysis.

These features are decision-support tools. They are not objective or final determinations of quality.

A review or score may reflect the information available to the system, the criteria applied, the design choices made by the user, and the limitations of the underlying AI model. A high score does not guarantee that content is accurate, effective, accessible, compliant, safe, or implementation-ready. A lower score does not necessarily mean that the content is unsuitable.

Users should consider Lira’s findings alongside professional judgment, subject-matter expertise, learner context, organizational requirements, and appropriate review processes.

Lira does not use personal information or AI-assisted features to make fully automated decisions that produce legal or similarly significant effects about users or other individuals.

Users may not use Lira Outputs, recommendations, scores, or evaluations as the sole basis for decisions that materially affect a learner, employee, applicant, educator, or other individual.

8. Translation and Language Adaptation

AI-assisted translation and language adaptation may contain errors in meaning, terminology, tone, grammar, cultural context, readability, or instructional intent.

Users are responsible for reviewing translated or adapted content before use. Materials intended for publication, formal instruction, legal compliance, safety guidance, assessment, or high-impact use should be reviewed by a qualified speaker or language specialist.

Lira does not guarantee that translated content meets local language standards, regional conventions, accessibility requirements, or organizational terminology requirements.

9. User Content and AI-Generated Outputs

Users retain ownership of the content they upload, enter, or create in Lira Studio.

As between the user and Lira, Lira does not claim ownership of user-specific AI outputs generated through the Services, except for Lira materials that may be included in or reflected by an output. User rights in AI-generated material remain subject to applicable law, third-party rights, and the Terms and Conditions.

Lira retains all rights in its platform and proprietary materials, including its:

Software and interface.
Research synthesis and instructional guidance.
Readiness frameworks and calibration methods.
Taxonomies, libraries, templates, and structures.
Review methods and evaluation criteria.
System instructions, processes, and platform logic.
Branding and other intellectual property.

Users may use AI-assisted outputs for the professional and commercial purposes permitted by their Lira account or agreement. Users may not extract, copy, reconstruct, resell, or use Lira’s underlying system, proprietary guidance, or platform materials to create a competing product or service.

10. Outputs May Not Be Unique

AI systems may produce the same or similar content for different users, particularly when users submit similar instructions, source material, or project context.

Lira does not guarantee that an output is unique, exclusive, copyrightable, or free from similarity to material created for another user.

Users are responsible for checking whether an output is appropriate for publication, licensing, sale, trademark use, or any other purpose that requires originality or exclusive rights.

11. Copyright and Third-Party Rights

Users must have the right to submit content to Lira Studio and to direct Lira to process that content.

Users may not knowingly use Lira to reproduce, transform, distribute, or create derivative works from material in a way that violates copyright, trademark, privacy, confidentiality, contractual, or other legal rights.

AI-generated outputs may contain material that resembles existing language, ideas, structures, or content. Before publishing or commercially distributing an output, users are responsible for reviewing it for possible infringement, improper attribution, or unauthorized use.

Lira does not provide legal advice or guarantee that an output is eligible for copyright protection or does not infringe another party’s rights.

12. Bias, Fairness, and Learner Variation

Lira is designed to help users consider learner readiness, population needs, context, support, and variation. These features cannot account for every individual, identity, culture, community, language, disability, environment, or lived experience.

Users should review content for:

Stereotypes and unsupported assumptions.
Cultural and linguistic relevance.
Representation and exclusion.
Accessibility and accommodation needs.
Differences across learners and environments.
Bias in source material, prompts, generated content, and evaluation criteria.
Unintended consequences for particular groups.

AI should not be used to label a learner, diagnose a condition, determine ability, predict future success, or make high-impact educational or employment decisions.

13. Transparency About AI Use

Lira identifies features that generate, revise, translate, analyze, or review content using artificial intelligence.

Users are responsible for providing additional disclosure when required by law, contract, organizational policy, publishing requirements, academic rules, or professional standards.

Depending on the context, users may need to disclose that:

AI assisted in creating or revising content.
A translation was AI-assisted.
An analysis, review, recommendation, or score was generated with AI.
A person reviewed and approved the final material.

Lira may include AI-use information, labels, notices, or metadata within the platform or exported materials. Users may not remove or alter a required disclosure when doing so would be deceptive or unlawful.

Certain transparency obligations under Article 50 of the European Union’s AI Act apply beginning August 2, 2026. These include requirements concerning notice of AI interaction and the identification of certain AI-generated or altered content. The exact obligation depends on the system, output, and use case. Additional information is available in the European Commission’s guidelines on AI transparency obligations.

14. Prohibited Uses

Users may not use Lira Studio or its AI-assisted features to:

Make fully automated decisions that produce legal or similarly significant effects about users or other individuals.
Diagnose medical, psychological, developmental, or learning conditions.
Determine special education eligibility, accommodations, discipline, grades, admissions, employment, or access to services without appropriate human review and lawful procedures.
Generate unlawful, discriminatory, deceptive, harassing, abusive, or harmful content.
Create content intended to exploit, manipulate, or endanger children or other vulnerable individuals.
Impersonate another person or misrepresent AI-generated content as independently verified human work when disclosure is required.
Submit content the user does not have permission to use.
Circumvent safeguards, access controls, usage limits, or platform restrictions.
Reverse engineer Lira’s AI instructions, proprietary guidance, evaluation methods, or platform logic.
Use Lira to develop or improve a competing model, platform, dataset, evaluation system, or commercial service without written permission.

Additional prohibited activities are described in the Terms and Conditions.

15. Changes to Models, Providers, and Features

AI systems change over time. Lira may add, remove, update, or replace AI models, service providers, instructions, evaluation methods, safeguards, or features.

These changes may affect the style, speed, quality, consistency, or behavior of outputs. Lira does not guarantee that the same instruction will always produce the same result.

We will update this Policy and provide appropriate notice when a change materially affects user rights, privacy, or how project content is processed.

16. Reporting a Concern

Users may report inaccurate, unsafe, biased, infringing, or otherwise concerning AI-generated content by contacting:

Lira Learning LLC
Email: support@liralearning.com

Please do not include student information or other sensitive personal information in a report. Include the project name, feature used, a description of the concern, and a non-sensitive example when possible.

17. Changes to This Policy

We may update this Policy as Lira Studio develops, our AI providers change, or legal and technical requirements evolve.

We will post the revised Policy with a new effective date. We will notify registered users by email before material changes take effect when appropriate or required.

Continued use of the Services after an updated Policy takes effect is subject to the updated Policy and the Terms and Conditions.