This section records the production context and the generative AI tools, models, and pipeline components used. Provide accurate, current information. If a detail is unknown at this stage, note it as pending rather than leaving it blank.
Production or project title
Production code or internal reference
Production format
Feature film
Television series or episode
Short film
Commercial or branded content
Music video
Documentary
Game cinematic or real-time sequence
Virtual production or live event
Other
Disciplines where generative AI is applied on this production
Visual effects and compositing
Concept art and previs
Scriptwriting and story development
Character, dialogue, or outline drafting
Score and music composition
Sound design and audio elements
Voice synthesis or dubbing
Marketing artwork and trailers
Subtitles and localisation
Editorial and assembly support
Other
Production company or studio
Department or unit submitting this form
Submitting supervisor or department lead
Contact email
Contact phone or preferred messaging handle
Planned start date for AI-related work
Target delivery date for AI-related material
Is any AI-related work performed by an external vendor, contractor, or freelance contributor?
Is any generative AI model fine-tuned, trained, or adapted on project-specific material?
List every generative tool, model, or service that contributes material to the production, including hosted services, locally installed software, plug-ins, and in-house systems.
AI Tool Technical Stack Register
Tool or model name | Version or build | Developer or vendor | Deployment (hosted service, local install, or in-house) | Primary use on this production | Provider documentation reviewed? | Operator or responsible person | Outputs retained? | |
|---|---|---|---|---|---|---|---|---|
Text-to-image generation model | Build 3.2 | Independent software developer | Hosted service | Concept art and environment exploration | Concept art lead | |||
Language model for drafting | Build 1.8 | Independent software developer | Local install | Scene outline and dialogue variants | Writing team | |||
Audio synthesis model | Build 2.0 | In-house research group | In-house system | Temp music cues and texture layers | Music editor | |||
Level of human creative oversight applied to AI output
Fully human-directed with iterative review
AI-assisted with substantial human editing
AI-generated with light touch-ups
AI-generated output used largely as-is
Varies by department or shot
Do any AI-generated elements appear in the final deliverable without further human modification?
Approximate number of AI-generated shots or assets expected in the final deliverable
Review checkpoints applied before AI output is approved
Department lead review
Editorial review
Continuity and story review
Rights or clearance review
Client or studio approval
Performer or representative approval
External quality assurance
None formally defined
Other
Technical documentation, tool inventory, or pipeline notes (optional)
Prompt and generation records support traceability, reproducibility, and internal review. Record the instruction or brief given to the tool, the parameters used, the resulting asset, and the person who approved it.
Have all prompts, parameters, and generation sessions been logged and retained?
A complete prompt and generation log is attached to this submission.
Prompt and Asset Generation Log
Prompt or session ID | Prompt or instruction summary | Tool and version | Generation date | Iterations generated | Iterations retained | Retention rate (%) | Used in the final deliverable? | |
|---|---|---|---|---|---|---|---|---|
P-001 | Wide establishing view of a rain-soaked harbour district at dusk, muted colour palette, no text or signage | Text-to-image model, build 3.2 | 3/12/2027 | 42 | 6 | 14.29 | ||
P-002 | Dialogue variant for the interrogation scene, two speakers, restrained tone, under 120 words | Language model, build 1.8 | 3/15/2027 | 18 | 4 | 22.22 | ||
Prompting and generation techniques used
Text-to-image generation
Image-to-image variation
Inpainting or outpainting
Style or aesthetic transfer
Reference-image conditioning
Fine-tuning or adapter training
Text drafting and rewriting
Idea and outline generation
Dialogue or lyric drafting
Music or audio synthesis
Voice synthesis
Motion or animation generation
Upscaling, restoration, or cleanup
Other
Did any prompt name, describe, or depict a real, identifiable person or performer?
Did any prompt reference an existing protected work, series, franchise, or the name of a living artist, writer, or composer?
Are prompt records versioned and time-stamped?
Yes – fully versioned and time-stamped
Partially – some records are versioned
No – records are informal or unversioned
Not sure
Total number of generation iterations recorded for the project so far
Estimated share of AI-generated material retained in the final deliverable (%)
Were content filters, negative prompts, or safety settings used to reduce unintended output?
Can the generation process be reproduced from the retained records?
Notes on unusual generation events, rejected outputs, manual corrections, or unexpected artefacts
Prompt log export, generation history, or sample prompt sheets
This section records what is known about the origin of the data behind the tools used, and the checks carried out to reduce the chance that generated material resembles existing protected works. Where information is limited or unavailable, state that plainly. Nothing in this section constitutes legal advice or a determination of rights.
What is known about the training data behind the tools used on this production?
Disclosed publicly in provider documentation
Provided on request under a written agreement
Limited or partial information available
Not disclosed by the provider
Custom or in-house dataset assembled for this project
Not yet reviewed
Steps taken to review dataset provenance
Reviewed provider documentation
Requested written information on data sources
Checked for opt-out or exclusion mechanisms
Reviewed the tool's licence and usage terms
Searched for known rights claims or disputes
Consulted an internal rights or clearance team
Consulted an external specialist adviser
Commissioned an independent review
No steps taken yet
Other
Has the tool provider made written statements about how training data was sourced or licensed?
Are you aware of any training material that may be subject to a third-party rights claim?
Assessed risk that generated output could resemble existing protected material
Assessed as low
Assessed as moderate
Assessed as elevated
Not yet assessed
Unable to assess with available information
Similarity and originality checks performed
Reverse image search
Visual similarity review by the art department
Text similarity or plagiarism scan for scripts
Music recognition or audio fingerprinting
Comparison against a reference library
Comparison against the studio's existing asset library
External clearance or title search
Specialist legal review
No checks performed
Other
Has each AI-generated asset been reviewed by a named person before use?
Does any generated material rely on an open or public licence that imposes conditions?
Asset Provenance Register
Asset ID | Description | Generation tool | Provenance status | Licence or permission basis | Similarity check completed? | Reviewed by | Notes | |
|---|---|---|---|---|---|---|---|---|
A-101 | Harbour district matte painting, final version | Text-to-image model, build 3.2 | Provider documentation reviewed | Tool output terms accepted by production | Concept art lead | Compared against reference library; no close matches noted | ||
A-118 | Temp music cue for the closing sequence | Audio synthesis model, build 2.0 | In-house dataset, sources documented | In-house contributor agreement | Music editor | Fingerprinting check scheduled before picture lock | ||
Known claims, disputes, challenges, or internal concerns relating to AI-generated material on this production
The provenance information in this section is accurate and complete to the best of my knowledge.
This section covers any synthetic recreation, imitation, or alteration of a real person's likeness, voice, movement, or performance style, and any use of archival or historical performer material. Record what was agreed, with whom, and on what basis. This form does not determine whether any arrangement is enforceable in any particular place.
Does the production use any synthetic version of a real person's likeness, voice, or performance?
Types of synthetic performance or likeness use on this production
Face replacement or de-ageing
Digital double
Voice cloning or replication
Synthetic voice for new dialogue
Lip-sync alteration or re-voicing
Body or pose synthesis
Performance style emulation
Synthetic background performers or crowds
Archival or historical performer material
Alteration of an existing recorded performance
Other
Has each affected performer, or their authorised representative, given written consent?
Scope of the consent obtained
This production only
This production plus specified future uses
A defined term, territory, and set of media
Consent is still under negotiation
Not applicable – no real performer is involved
Other
Information shared with the performer or representative before consent was given
Nature and extent of the synthetic use
Duration or term of use
Media, platforms, and formats
Territories
Compensation or other consideration
Credit and attribution
Approval and review rights
Withdrawal or termination options
Data storage and retention
Limits on editing or alteration
Context in which the material will appear
Other
Was the performer given an opportunity to review and approve the synthetic material?
Total consideration or compensation agreed for synthetic use (per performer or in total)
Agreed credit or attribution wording
Are there agreed limits on how the synthetic material may be edited, combined, or reused?
Does the production use material connected to a person who is deceased or otherwise unable to respond directly?
Were any real voices, likenesses, or performances used to train or fine-tune a model for this production?
Are synthetic performers or voices clearly identified in the production's internal records and credits plan?
Performer and Consent Register
Performer or artist name | Role or capacity | Type of synthetic use | Consent obtained? | Consent date | Term and territory | Representative contact | Notes | |
|---|---|---|---|---|---|---|---|---|
Sample entry – lead performer | Principal cast | Voice synthesis for additional dialogue | 2/18/2027 | This production only, worldwide | Agent contact on file | Approval required for each new line before final mix | ||
Sample entry – featured artist | Guest musical artist | Style emulation for a transitional cue | Under discussion | Management contact on file | No material to be used until written consent is received | |||
All synthetic performance material is recorded in the register above or in an attached schedule.
Consent forms, agreements, or representative correspondence
This final section records how rights in AI-assisted material are expected to be held, which documents have been reviewed, and what has been disclosed to other parties. Statements here reflect the submitter's understanding at the time of signing and are not a legal opinion or a guarantee that any material is free of third-party rights.
How are rights in AI-assisted material expected to be held?
Studio-owned under a work-for-hire style arrangement
Assigned to the studio by the contributor
Joint ownership between parties
Licensed from a third-party rights holder
Governed by the tool provider's output terms
Not yet determined
Documents and licences reviewed for this production
Tool terms of service
Model or output licence terms
Vendor or services agreement
Contributor and crew agreements
Performer consent documents
Music synchronisation and master licences
Stock or library asset licences
Open licence obligations
Location, property, or artwork releases
Confidentiality and non-disclosure agreements
None reviewed yet
Other
Do the tool's terms grant the studio the right to use outputs for this production's intended purposes?
Do any outputs carry attribution, share-alike, or similar obligations?
Has a specialist in intellectual property or contract terms reviewed the AI-related provisions?
Disclosures made to financiers, distributors, or insurers
Written disclosure of generative AI use
Disclosure of the tool and model list
Disclosure of synthetic performance use
Disclosure of dataset provenance status
Disclosure of prompt retention practices
Disclosure of known rights claims or disputes
No disclosures made yet
Other
Do any financing, distribution, or delivery agreements contain terms that affect AI-generated material?
Conditions or restrictions attached to the use of generated material
No additional conditions identified
Credit or attribution required
Use limited to specified media or territories
Use limited in time
No use in training of other models
No use in sensitive or high-risk contexts
Editing or alteration restricted
Approval required before release
Other
Summary of residual risks and the steps proposed to manage them
Overall clearance recommendation for AI-related material
Cleared for delivery
Cleared subject to listed conditions
Pending further review
Not cleared at this time
Referred for specialist advice
Recommended date for the next review or re-check
Rights and Licence Register
Item or asset | Rights holder | Licence type | Territory | Term | Expiry or renewal date | Document reference | Confirmed? | |
|---|---|---|---|---|---|---|---|---|
Generated concept artwork | Production studio | Contributor assignment | Worldwide | Perpetual for this production | 1/30/2027 | AGR-2027-004 | ||
Synthetic voice material | Performer and production studio | Consent and licence | Worldwide | This production only | 2/18/2027 | CON-2027-011 | ||
Temp music cue | In-house composer | In-house agreement | Worldwide | Under review | MUS-2027-002 | |||
The authorisations described in this section have been obtained or are being pursued as stated.
Declaration: this form records an internal review of generative AI-related material for the named production. It is not legal advice, it does not represent approval by any external authority, and it does not guarantee that any material is free of third-party rights.
Authorised representative signature
Name and title of signatory
Date of authorisation
I confirm that the information in this form is complete and accurate to the best of my knowledge.
Analysis for Generative AI Clearance Form for Production Teams
Important Note:
This analysis provides strategic insights to help you get the most from your form's submission data for powerful follow-up actions and better outcomes.
AI-Generated Content: This template and its accompanying guidance are AI-generated to provide a practical starting point for your form-building needs. Please review and customise the content to ensure it is accurate, appropriate and suitable for your intended use.
This guide is provided for general informational and planning purposes only. It does not constitute professional, legal, financial, medical, employment, privacy, regulatory, or advice of any other kind, nor does it certify that this template is suitable or compliant for any particular jurisdiction or purpose.
Laws, regulations, industry practices and business requirements vary. You are responsible for reviewing and customising the template to suit your intended use and obtaining advice from appropriately qualified professionals where necessary.
Please delete this section before publishing your form. See our Terms of Service for further information.
This Generative AI Clearance Form for Production Teams is a well-structured, five-section document designed to capture how generative AI tools are used across a production in visual effects, scriptwriting, or score composition. The form covers production details and tool stacks, prompt and asset generation logs, dataset provenance and copyright checks, synthetic performer clearances, and studio-level rights and licensing authorisations. Its breadth is a clear strength: it moves from operational context (who, what, when) through technical traceability (how outputs were generated) to rights and consent matters (who owns what, what was agreed), producing a comprehensive record. The form also includes sample table entries that help respondents understand the expected level of detail. Areas for improvement include the overall length of the form, which may challenge completion rates if filled in a single sitting, and the large number of mandatory fields, which could be streamlined through more aggressive use of conditional logic.
Production or project title serves as the primary identifier for the entire submission. It anchors all subsequent records — prompt logs, asset registers, consent documents — to a single named production, which is essential for internal record-keeping and cross-referencing.
Production code or internal reference is an optional field that supports integration with existing production management systems. Including it as optional is a sensible choice, since not every team will use formal project codes.
Production format captures the delivery context (feature film, series, commercial, etc.), which influences the scale, rights considerations, and review processes that may apply. The conditional follow-up for 'Other' ensures unusual formats are documented.
Disciplines where generative AI is applied on this production is a multiple-choice field that maps AI usage across the production pipeline. This is operationally valuable because it helps reviewers understand the scope of AI involvement at a glance and identify which departments may require additional scrutiny.
Production company or studio and Submitting supervisor or department lead establish accountability. Knowing who is responsible for the submission and which organisation stands behind it supports follow-up communication and internal governance.
Contact email and Contact phone or preferred messaging handle provide communication channels. The email field is mandatory, reflecting its role as the primary digital touchpoint, while the phone field is optional — a proportionate approach.
Planned start date for AI-related work and Target delivery date for AI-related material establish a timeline. The start date being mandatory helps reviewers understand when AI-related activities begin, which may be relevant for scheduling reviews or approvals.
The two yes/no questions — Is any AI-related work performed by an external vendor, contractor, or freelance contributor? and Is any generative AI model fine-tuned, trained, or adapted on project-specific material? — are well-designed gate questions. Their conditional follow-ups only appear when relevant, reducing unnecessary data entry for teams that use only off-the-shelf tools in-house.
The AI Tool Technical Stack Register table is a strong design element. It captures tool names, versions, deployment types, operators, and whether provider documentation has been reviewed and outputs retained. The sample rows demonstrate the expected granularity. This register is central to traceability: it documents exactly which tools contributed material and who operated them.
Level of human creative oversight applied to AI output captures the degree of human involvement, with a conditional follow-up for the least human-directed option. This is a pragmatic design that prompts additional explanation where oversight is minimal.
Do any AI-generated elements appear in the final deliverable without further human modification? is another gate question that flags potentially higher-risk material. The follow-up asks for specific identification of shots or assets, which supports targeted review.
Approximate number of AI-generated shots or assets expected in the final deliverable provides a quantitative estimate useful for workload planning and risk assessment.
Review checkpoints applied before AI output is approved captures the internal review process. The option 'None formally defined' with a follow-up about how gaps will be addressed is a thoughtful touch that encourages process improvement.
Have all prompts, parameters, and generation sessions been logged and retained? is a foundational traceability question. Its dual follow-up (where logs are stored if yes, or what is missing and how it will be remediated if no) ensures both positive and negative responses yield actionable information.
The Prompt and Asset Generation Log table is one of the form's strongest elements. It includes a calculated Retention rate (%) column using a formula, which automates a useful metric without requiring manual calculation. The sample entries show realistic prompt summaries, iteration counts, and final-deliverable flags.
Did any prompt name, describe, or depict a real, identifiable person or performer? and Did any prompt reference an existing protected work, series, franchise, or the name of a living artist, writer, or composer? are critical screening questions. Their conditional follow-ups ask for specifics — names, consent references, rejected outputs — which supports targeted rights review.
Are prompt records versioned and time-stamped? assesses record-keeping maturity. The graduated options (fully, partially, not, not sure) with tailored follow-ups for each are well-designed and avoid forcing a binary answer.
Can the generation process be reproduced from the retained records? is a key operational question. The dual follow-up for yes and no responses ensures that both known limits and remediation plans are documented.
The optional fields on content filters, total iterations, and estimated share of AI material in the final deliverable provide useful context without overburdening the respondent.
What is known about the training data behind the tools used on this production? is a nuanced single-choice question with follow-ups tailored to each option. This is effective design: it acknowledges that information may be limited and asks respondents to state what they know plainly, without implying that full disclosure is always achievable.
Steps taken to review dataset provenance captures the diligence steps undertaken. The inclusion of 'No steps taken yet' with a follow-up about plans is honest and practical.
Are you aware of any training material that may be subject to a third-party rights claim? is a direct risk-screening question. The 'no' follow-up asking for the basis of the conclusion is a thoughtful addition that encourages documented reasoning rather than assumptions.
Assessed risk that generated output could resemble existing protected material uses a graduated scale (low to unable to assess) with appropriate follow-ups. This avoids forcing a false sense of certainty.
Has each AI-generated asset been reviewed by a named person before use? reinforces individual accountability. The follow-ups ask how reviews are recorded or what the plan is for unreviewed assets.
The Asset Provenance Register table mirrors the tool stack register in structure but focuses on individual assets, their provenance status, licence basis, and similarity check completion. This per-asset granularity is operationally valuable.
The closing checkbox — The provenance information in this section is accurate and complete to the best of my knowledge — serves as a declaration point, which is a common and practical pattern in clearance workflows.
Does the production use any synthetic version of a real person's likeness, voice, or performance? is the section's primary gate question. The 'no' follow-up includes a helpful paragraph directing respondents to Section 5 while noting that remaining questions may still be useful — a considerate UX choice.
Has each affected performer, or their authorised representative, given written consent? is a central consent-tracking question. The follow-ups (upload consent documents if yes, or explain status and steps if no) ensure that both compliant and in-progress situations are documented.
Scope of the consent obtained captures the boundaries of what was agreed — production-only, future uses, defined terms, or still under negotiation. This is important for operational planning.
Was the performer given an opportunity to review and approve the synthetic material? tracks whether performers had review rights. The follow-ups ask for approval records or alternative safeguards.
Does the production use material connected to a person who is deceased or otherwise unable to respond directly? addresses a sensitive scenario. The follow-up asks for details of who was approached and the outcome, which is proportionate and respectful.
Were any real voices, likenesses, or performances used to train or fine-tune a model for this production? captures a specific and potentially significant use case. The follow-up asks for the material, ownership, and basis of use.
The Performer and Consent Register table provides per-performer tracking with consent dates, terms, territories, and representative contacts. The sample entries demonstrate realistic scenarios including one where consent is still pending.
How are rights in AI-assisted material expected to be held? captures the anticipated ownership model. The graduated options (studio-owned, assigned, joint, licensed, governed by tool terms, not yet determined) cover common arrangements without being exhaustive.
Do the tool's terms grant the studio the right to use outputs for this production's intended purposes? is a practical operational question. The follow-ups ask for conditions or restrictions, which supports downstream planning.
Has a specialist in intellectual property or contract terms reviewed the AI-related provisions? tracks whether expert review has occurred. The follow-ups ask who reviewed and when, or who will review and by when — a balanced approach.
Overall clearance recommendation for AI-related material is the form's culminating decision point. The graduated options (cleared, cleared with conditions, pending, not cleared, referred) with tailored follow-ups provide a clear and actionable outcome.
The closing elements — declaration paragraph, signature, name and title, date, and accuracy confirmation checkbox — follow a standard authorisation pattern that provides a clear record of who approved the submission and when.
This form collects a substantial volume of structured data across tool inventories, prompt logs, asset registers, consent records, and rights documentation. The use of tables with sample entries promotes consistency in data entry. The calculated retention-rate column in the prompt log reduces manual errors. From a privacy perspective, the form collects contact details, performer names and consent information, and compensation figures — all of which warrant appropriate handling and access controls within the production organisation. The form's declaration paragraph appropriately notes that it is an internal record and not a legal opinion.
The form is logically ordered, moving from broad context to specific rights matters. Each section opens with a clarifying paragraph, which helps respondents understand expectations before answering. Conditional follow-ups reduce unnecessary data entry. However, the form is extensive — five sections with numerous questions, tables, and file uploads — which may require significant time to complete. Breaking the form into a multi-step or save-and-resume process could improve completion rates. The sample table entries are helpful, but respondents unfamiliar with clearance workflows may still benefit from brief tooltips or examples for complex fields such as dataset provenance status or consent scope.
Mandatory Question Analysis for Generative AI Clearance Form for Production Teams
Important Note:
This section explains why certain fields in this form are marked as required and provides general guidance on how these requirements may support the collection of complete, accurate and relevant information.
Required fields and information-collection obligations may vary depending on your industry, jurisdiction, business operations and intended use of the form. A field marked as required in this template does not necessarily mean that collecting the information is legally mandatory.
Please review each required field to determine whether it is necessary, appropriate and lawful for your particular purpose. Consider applicable privacy, data protection, consent and industry-specific requirements, and consult an appropriately qualified legal, privacy or compliance professional where necessary.
Please delete this section before publishing your form. See our Terms of Service for further information.
Production or project title
Requiring this field ensures every submission is clearly identified and can be linked to the correct production. A title serves as the primary reference across all registers and documents attached to the form. Without it, downstream record-keeping and cross-referencing become unreliable.
Production format
Requiring the production format helps reviewers understand the delivery context and scale of the project. Different formats may carry different review workflows or stakeholder expectations. Making this mandatory ensures the submission can be routed appropriately.
Disciplines where generative AI is applied on this production
Requiring this field captures the scope of AI involvement across departments. It helps reviewers identify which areas of the production may need additional scrutiny or documentation. A complete picture of AI usage supports more targeted review processes.
Production company or studio
Requiring the production company or studio establishes the organisational context for the submission. It supports accountability and ensures the form can be filed under the correct entity in internal records.
Submitting supervisor or department lead
Requiring the name of the submitting supervisor or department lead creates a clear point of accountability. It ensures there is a named person who can be contacted for clarification and who takes responsibility for the accuracy of the submission.
Contact email
Requiring a contact email supports ongoing communication about the submission, including follow-up questions, review scheduling, and notification of outcomes. It serves as the primary digital channel for correspondence related to this form.
Planned start date for AI-related work
Requiring the planned start date helps reviewers understand when AI-related activities are expected to begin. This may be relevant for scheduling reviews, aligning with production milestones, or ensuring that documentation is in place before work commences.
Is any AI-related work performed by an external vendor, contractor, or freelance contributor?
Requiring this gate question ensures that third-party involvement in AI-related work is identified early. External contributors may introduce additional contractual or operational considerations, and knowing their involvement helps reviewers assess the full scope of the production's AI pipeline.
Is any generative AI model fine-tuned, trained, or adapted on project-specific material?
Requiring this question flags whether custom model adaptation is taking place, which may involve additional data handling, contributor agreements, or provenance considerations. Identifying this early supports more thorough documentation.
Level of human creative oversight applied to AI output
Requiring this field captures the degree of human involvement in shaping AI-generated material. This information helps reviewers understand the creative process and may inform how outputs are assessed or documented. The necessity of this field depends on the form creator's review framework.
Do any AI-generated elements appear in the final deliverable without further human modification?
Requiring this question identifies material that may warrant closer review, since unmodified AI output may carry different characteristics than human-edited material. The follow-up asks for specific identification, supporting targeted assessment.
Have all prompts, parameters, and generation sessions been logged and retained?
Requiring this question establishes whether a traceability record exists. Prompt logs support reproducibility, internal review, and documentation of the creative process. Requiring an answer ensures the status of record-keeping is explicitly stated rather than assumed.
Did any prompt name, describe, or depict a real, identifiable person or performer?
Requiring this screening question ensures that prompts involving real individuals are flagged. This may trigger follow-up on consent or clearance matters documented elsewhere in the form. Requiring an answer prevents this important detail from being overlooked.
Did any prompt reference an existing protected work, series, franchise, or the name of a living artist, writer, or composer?
Requiring this question identifies prompts that reference existing creative works or named creators. This supports awareness of potential similarity or attribution considerations. The follow-up asks for specifics, which aids targeted review.
Can the generation process be reproduced from the retained records?
Requiring this question assesses whether the generation workflow is documented well enough to be replicated or reviewed. Reproducibility supports internal quality assurance and may be relevant to the production's own record-keeping standards. The dual follow-up ensures both strengths and gaps are captured.
What is known about the training data behind the tools used on this production?
Requiring this question ensures the production team explicitly states the level of information available about training data. This supports informed decision-making and documentation of any limitations. The graduated options acknowledge that full information may not always be available.
Are you aware of any training material that may be subject to a third-party rights claim?
Requiring this question ensures awareness of potential rights claims is explicitly recorded. The follow-up for 'no' responses asks for the basis of the conclusion, which encourages documented reasoning. This supports the production's own risk awareness.
Assessed risk that generated output could resemble existing protected material
Requiring this field captures the production team's assessment of similarity risk. The graduated scale avoids forcing certainty where information is limited. This assessment may inform downstream review priorities.
Has each AI-generated asset been reviewed by a named person before use?
Requiring this question reinforces individual accountability for asset review. It supports the production's internal quality and rights-check processes. The follow-ups ask for review records or plans for unreviewed assets, which promotes completeness.
The provenance information in this section is accurate and complete to the best of my knowledge.
Requiring this declaration checkbox ensures the submitter explicitly confirms the accuracy of the provenance information. It serves as a record of attestation and encourages careful review before submission.
Does the production use any synthetic version of a real person's likeness, voice, or performance?
Requiring this gate question ensures that synthetic performer use is explicitly identified. This determines whether the remaining questions in the section are relevant and helps reviewers understand the scope of synthetic material on the production.
Has each affected performer, or their authorised representative, given written consent?
Requiring this question ensures consent status is explicitly recorded for each affected performer. The follow-ups capture either the consent documentation or the current status and remediation steps. This supports the production's own consent tracking processes.
Scope of the consent obtained
Requiring this field captures the boundaries of what was agreed — whether consent covers this production only, future uses, defined terms, or is still under negotiation. This information supports operational planning and helps avoid use beyond agreed limits.
Was the performer given an opportunity to review and approve the synthetic material?
Requiring this question records whether performers had review rights over synthetic material. The follow-ups ask for approval records or alternative safeguards. This supports transparent consent practices and internal documentation.
Does the production use material connected to a person who is deceased or otherwise unable to respond directly?
Requiring this question ensures that material connected to persons who cannot respond directly is identified. The follow-up asks for details of who was approached and the outcome, which supports respectful and thorough record-keeping.
Were any real voices, likenesses, or performances used to train or fine-tune a model for this production?
Requiring this question captures whether real performer material was used in model training or adaptation. The follow-up asks for details of the material, ownership, and basis of use, which supports documentation of this specific activity.
How are rights in AI-assisted material expected to be held?
Requiring this field captures the anticipated ownership model for AI-assisted material. This is a central question for the production's rights documentation and helps reviewers understand the expected chain of title. The necessity of specific options may depend on the form creator's contractual framework.
Do the tool's terms grant the studio the right to use outputs for this production's intended purposes?
Requiring this question ensures the production team has considered whether tool terms permit the intended use of outputs. The follow-ups capture conditions or restrictions, which supports downstream planning and documentation.
Has a specialist in intellectual property or contract terms reviewed the AI-related provisions?
Requiring this question records whether expert review of AI-related provisions has taken place. The follow-ups ask who reviewed and when, or who will review and by when. This supports the production's own review workflow, though the necessity of specialist review may vary by context.
Do any financing, distribution, or delivery agreements contain terms that affect AI-generated material?
Requiring this question ensures that contractual terms affecting AI-generated material are identified. The follow-up asks for a summary of relevant terms, which supports awareness of any obligations or restrictions that may apply.
Overall clearance recommendation for AI-related material
Requiring this field captures the form's culminating assessment. The graduated options provide a clear outcome — cleared, cleared with conditions, pending, not cleared, or referred. The follow-ups for each option ensure the reasoning is documented.
The authorisations described in this section have been obtained or are being pursued as stated.
Requiring this declaration checkbox ensures the submitter explicitly confirms the status of authorisations. It serves as a record of attestation and encourages careful review before submission.
Authorised representative signature
Requiring a signature provides a formal record of authorisation. It identifies the individual who approved the submission and creates a clear accountability point.
Name and title of signatory
Requiring the name and title of the signatory ensures the authorised representative is clearly identified in text form, complementing the signature. This supports record-keeping and future reference.
Date of authorisation
Requiring the date of authorisation establishes when the submission was approved. This supports timeline tracking and may be relevant for aligning with production milestones or review schedules.
I confirm that the information in this form is complete and accurate to the best of my knowledge.
Requiring this final confirmation checkbox ensures the submitter explicitly attests to the accuracy and completeness of the entire form. It serves as a closing declaration and encourages thorough review before submission.
This form contains a substantial number of mandatory fields — approximately 36 across five sections. While the depth of information collected is appropriate given the form's clearance purpose, the volume of required fields may affect completion rates, particularly if respondents are filling in the form under time pressure or without prior experience with clearance workflows.
Balancing data collection depth against completion rates: Consider whether all currently mandatory fields are essential at the point of first submission, or whether some could be made optional and flagged for follow-up review. For example, fields that capture contextual detail (such as the level of human creative oversight or the scope of consent) might be made optional if the production team's primary goal is to capture the core clearance decisions. Alternatively, a phased approach — where a shorter initial submission is followed by a more detailed supplement — could help maintain data quality while reducing initial friction.
Conditional logic: The form already uses conditional follow-ups effectively, particularly in Sections 1, 2, and 4. However, there may be opportunities to extend this further. For example, if the answer to 'Does the production use any synthetic version of a real person's likeness, voice, or performance?' is 'No', the subsequent mandatory questions in Section 4 about consent scope and performer review could be made non-mandatory or skipped entirely, since they may not be relevant. Similarly, in Section 5, if no external vendors are involved and no financing agreements exist, some mandatory questions about agreement terms could be conditionally relaxed.
Best practices for mandatory field selection and placement: Mandatory fields are most effective when they capture information that is essential for the form's core purpose and that cannot reasonably be inferred or obtained later. In this form, identity fields (production title, company, supervisor, contact email), key gate questions (external vendor involvement, synthetic performer use, consent status), and the final authorisation elements (signature, name, date, confirmation checkbox) are strong candidates for mandatory status. Contextual and descriptive fields — such as notes on unusual generation events, estimated share of AI material, or review checkpoint details — may be better served as optional, since they enrich the record but are not always critical at the point of submission.
Overarching observations: The mandatory fields are distributed across all five sections, which is appropriate given the form's comprehensive scope. However, the density of mandatory fields in Sections 4 and 5 is notable. These sections deal with sensitive matters (performer consent, rights authorisation) where completeness is important, but respondents may benefit from a clear indication of which fields are strictly required versus which are recommended. A visual distinction (such as an asterisk or colour coding) between 'required for submission' and 'recommended for completeness' could help respondents prioritise their effort while still encouraging thorough documentation. Additionally, the form's closing declaration and signature elements are appropriately mandatory, as they serve as the formal attestation point for the entire submission.