Civil Law And Autonomous Creative System Copyright Revenue Claims In Singapore .
1. Introduction
Autonomous Creative System Copyright Revenue Claims concern legal disputes arising when artificial intelligence (AI), autonomous software, generative models, robotic creative platforms, or other computer-based systems independently or collaboratively produce commercially valuable creative content.
Examples include:
AI-generated paintings, illustrations and digital artwork.
Autonomous music composition and sound production.
AI-generated films, animations and audiovisual content.
Automated literary works, novels, articles and scripts.
AI-generated advertising campaigns and commercial designs.
Autonomous video-game characters, environments and digital assets.
AI-generated photographs and multimedia productions.
Creative content distributed through online platforms and subscription services.
In Singapore, these developments raise important civil-law questions concerning copyright ownership, authorship, licensing, royalties, revenue distribution, contractual obligations, infringement and unjust enrichment.
The central legal question is:
Who is entitled to receive copyright revenue when an autonomous creative system produces commercially successful content?
Potential claimants include:
The human creator who provides creative instructions and makes substantive artistic decisions.
The developer or operator of the AI system.
The commissioning company or commercial platform.
The employer of the human creator.
The copyright owner of underlying material incorporated into the output.
A licensee or distributor entitled to contractual royalties.
Singapore's legal framework does not automatically recognise an AI system as an independent copyright owner. Consequently, disputes generally require examination of human authorship, originality, contractual arrangements and the commercial rights associated with the resulting work.
2. Legal Framework in Singapore
The principal legislation is the Copyright Act 2021 (Singapore).
The Act provides the framework for copyright protection, ownership, licensing, infringement and remedies.
A. Copyright protection and originality
Copyright protection generally requires an identifiable work falling within a protected statutory category.
The relevant categories include:
Literary works.
Dramatic works.
Musical works.
Artistic works.
An AI-generated output may potentially qualify as a protected work where the necessary legal requirements are satisfied.
However, the involvement of an AI system does not automatically establish copyright protection.
The claimant must demonstrate the necessary human intellectual contribution and originality under the applicable legal principles.
B. Human authorship
Human authorship is particularly important in Singapore.
In Asia Pacific Publishing Pte Ltd v Pioneers & Leaders (Publishers) Pte Ltd, the Court of Appeal emphasised the need to identify a human author and establish the connection between human intellectual effort and the creation of the work.
This creates important questions for autonomous creative systems:
Did a human exercise meaningful creative control?
Were the prompts sufficiently creative and specific?
Did the human select, arrange or substantially modify the output?
Was the final work generated predominantly through autonomous machine processes?
Can the relevant creative contribution be attributed to an identifiable person?
A person who merely presses a generation button may have difficulty establishing authorship of the resulting output.
C. Copyright ownership and commercialisation
Copyright ownership determines who can commercially exploit protected works.
The Act provides for ownership arrangements involving employment, commissioned works, assignments and licences.
A written agreement can establish:
Ownership of AI-assisted creative works.
Rights to reproduce and distribute outputs.
Licensing and sublicensing arrangements.
Revenue-sharing percentages.
Royalty payment schedules.
Rights to derivative works.
Restrictions on commercial exploitation.
Obligations to maintain revenue records.
IPOS recognises that assignments and licences are important mechanisms for commercialising copyright and generating revenue.
D. Copyright infringement
Section 146 of the Copyright Act 2021 establishes the general framework for infringement through unauthorised performance of acts comprised in copyright.
Section 153 provides for infringement actions by copyright owners and qualifying exclusive licensees.
An autonomous creative platform may therefore become involved in litigation where its operations result in unauthorised reproduction, adaptation, communication or distribution of protected material.
Importantly, the existence of AI-generated content does not itself establish infringement. The claimant must establish the relevant protected rights, unauthorised conduct and other necessary elements.
WIPO+2
3. Major Civil-Law Issues in Autonomous Creative Copyright Revenue Claims
A. Disputes over AI-generated copyright ownership
One of the most significant disputes concerns whether the person operating an autonomous creative system owns the copyright in the generated material.
For example, a Singapore-based digital artist uses an AI platform to generate hundreds of illustrations and sells them through a commercial marketplace.
The platform operator subsequently claims that it owns the illustrations because its software generated them.
The artist may argue that their creative instructions, selection, editing and arrangement establish sufficient human authorship.
The court would need to examine the actual creative process rather than simply accepting either party's assertion.
B. Revenue-sharing disputes
Autonomous creative platforms may generate revenue through:
Subscription fees.
Commercial licences.
Advertising.
Digital downloads.
Streaming.
Marketplace transactions.
Commercial reproduction.
Licensing to film and gaming companies.
Disputes arise when the platform operator, developer, creator and distributor disagree about their respective shares.
A creator may claim that the platform has improperly retained royalties or failed to disclose commercial transactions.
Possible civil remedies include contractual damages, an account of profits where legally available, declarations of rights and appropriate injunctions.
C. AI training-data copyright claims
An autonomous creative system may be trained using copyrighted books, photographs, musical recordings, paintings and other creative material.
Original copyright owners may argue that their material was reproduced or commercially exploited without permission.
Potential disputes include:
Unauthorised reproduction during training.
Commercial use of protected source material.
Outputs substantially reproducing protected expression.
Failure to comply with licensing restrictions.
Disputes concerning attribution and royalty payments.
Singapore's computational data analysis exception is relevant, but it is subject to statutory conditions and restrictions. It should not be treated as unrestricted permission to commercially exploit copyrighted training material or generated outputs.
D. Disputes involving autonomous music production
AI music systems may generate melodies, instrumental compositions, arrangements and sound recordings.
Different copyright interests must be distinguished:
Copyright in the musical composition.
Copyright in lyrics.
Copyright in sound recordings.
Rights in performances.
Contractual royalty entitlements.
A person may own or control one category of rights without owning the others.
Consequently, revenue disputes can involve several different copyright owners and licensing arrangements.
E. Platform accountability
Online creative platforms may distribute AI-generated material to international customers.
A platform could face claims involving:
Unauthorised commercial distribution.
Breach of licensing agreements.
Misrepresentation of copyright ownership.
Incorrect royalty calculations.
Failure to account for licensing income.
Unauthorised sublicensing.
Contractual restrictions on commercial use.
The legal responsibility of each participant depends on the actual conduct, contractual arrangements and applicable statutory provisions.
F. Unjust enrichment and restitution
A developer or commercial platform may receive financial benefits from creative content while another party claims entitlement to that income.
The claimant may seek restitution or other appropriate civil remedies where the elements of an established cause of action are satisfied.
However, receiving revenue from AI-generated content does not automatically constitute unjust enrichment. The claimant must establish a legally recognised basis for recovery.
4. Important Case Laws and Judicial Principles
Important clarification: Singapore has no reported appellate decision directly establishing a comprehensive copyright-royalty regime for autonomous generative AI outputs. The following Singapore decisions and persuasive foreign authorities provide relevant principles. Foreign decisions are not binding on Singapore courts.
Case 1. Asia Pacific Publishing Pte Ltd v Pioneers & Leaders (Publishers) Pte Ltd [2011] 4 SLR 381
Singapore Court of Appeal
Facts and legal issue:
The dispute concerned copyright protection in newspaper advertising materials and whether the claimant could establish the necessary authorship and originality.
The Court of Appeal examined the relationship between authorship, originality and human intellectual contribution.
Judicial principle:
The identification of a human author is fundamental to establishing an original copyright work under Singapore's traditional copyright framework.
A corporation may own copyright without necessarily being the original author.
Relevance to autonomous creative systems:
This decision is particularly important where an AI system independently generates artistic or literary content.
A company cannot simply assert that it owns copyright because it developed or operated the software.
It must establish an appropriate legal basis for ownership, including identifiable human authorship where required, employment arrangements or valid assignment.
Application to revenue claims:
A human creator may challenge a platform's assertion that all revenue from AI-generated content belongs to the platform.
Case 2. Global Yellow Pages Ltd v Promedia Directories Pte Ltd [2016] 2 SLR 165
Singapore Court of Appeal
Facts and legal issue:
The case concerned copyright protection in telephone-directory compilations and the requirement of originality.
The Court considered the nature of intellectual creation and the extent of human effort necessary for copyright protection.
Judicial principle:
Copyright originality requires intellectual creation attributable to the author. Mere labour, expense or mechanical effort is not necessarily sufficient.
Relevance to autonomous creative systems:
An AI system may produce large volumes of content through automated processes.
However, the volume of generated material or computational effort does not automatically establish copyright protection.
The claimant must establish the necessary creative contribution.
Application to revenue claims:
A platform cannot assume that every automatically generated image, article or musical composition creates an independently enforceable copyright revenue stream.
Case 3. I-Admin (Singapore) Pte Ltd v Hong Ying Ting [2020] 1 SLR 1130
Singapore High Court
Facts and legal issue:
The litigation concerned confidential business information and the legal protection of commercially valuable information.
The court examined the protection of confidential information and the circumstances in which misuse may give rise to liability.
Judicial principle:
Confidential business information can receive legal protection independently of copyright where the requirements of the law of confidence are established.
Relevance to autonomous creative systems:
AI creative companies may possess valuable confidential information, including:
Proprietary model architecture.
Unpublished creative projects.
Commercial prompts and workflows.
Customer information.
Licensing databases.
Unreleased creative outputs.
Application to revenue claims:
A dispute over confidential creative assets may provide a separate cause of action even where copyright ownership in an AI-generated output is uncertain.
Case 4. RecordTV Pte Ltd v MediaCorp TV Singapore Pte Ltd [2010] 2 SLR 830
Singapore Court of Appeal
Facts and legal issue:
The dispute concerned an internet-based television recording service and whether the service provider had authorised copyright infringement.
The court considered the relationship between technological services, user conduct and authorisation liability.
Judicial principle:
The provision of technology does not automatically establish authorisation of copyright infringement. The actual conduct and degree of involvement of the service provider must be considered.
Relevance to autonomous creative systems:
An AI platform may provide tools that enable users to generate, reproduce or distribute creative material.
The platform's liability cannot be determined solely by establishing that it supplied the technology.
Its role in the infringing conduct, knowledge, control and relevant statutory provisions must be examined.
Application to revenue claims:
This case is relevant when a copyright owner seeks to hold a creative platform responsible for unauthorised commercial exploitation carried out through its system.
Case 5. Virtual Map (Singapore) Pte Ltd v Singapore Land Authority [2008] SGHC 42
Singapore High Court
Facts and legal issue:
The dispute concerned copyright in geographical mapping materials and the use of mapping information.
It involved questions about copyright protection and the exploitation of commercially valuable information.
Judicial principle:
Copyright protection must be examined by reference to the protected expression and the relevant rights in the underlying material.
Relevance to autonomous creative systems:
AI systems may generate maps, illustrations, architectural visualisations and other commercially valuable digital works.
The case helps illustrate why a claimant must identify the protected material and demonstrate the relevant infringement rather than claiming ownership over underlying facts or information generally.
Application to revenue claims:
A creator or company may pursue appropriate claims where another commercial platform reproduces protected digital creative material without authorisation.
Case 6. Nova Productions Ltd v Mazooma Games Ltd [2007] EWCA Civ 219
Persuasive English authority
Facts and legal issue:
The dispute concerned computer-generated images produced in the context of a computer game.
The court considered who should be treated as the author of computer-generated artistic material.
Judicial principle:
The court examined the role of the human programmer and distinguished between the person creating the program and the person merely operating the game.
Relevance to autonomous creative systems:
This decision is important when determining whether a developer, user or creative operator should receive authorship recognition.
A software developer does not automatically become the author of every output produced by the software.
Application to revenue claims:
The case provides a useful basis for analysing whether the creator of an AI platform or the person making substantive creative decisions should receive copyright-related commercial benefits.
Case 7. CCH Canadian Ltd v Law Society of Upper Canada [2004] 1 SCR 339
Persuasive Canadian authority
Facts and legal issue:
The dispute involved copyright in legal publications and materials reproduced through a legal research service.
The court considered originality and the degree of intellectual effort necessary to establish copyright protection.
Judicial principle:
Originality requires an exercise of skill and judgment that is not purely mechanical.
Relevance to autonomous creative systems:
This principle is useful when examining human participation in AI-assisted creation.
For example, a human editor who substantially selects, arranges and modifies AI-generated content may have a stronger argument for copyright protection than someone who merely accepts an automatically generated result.
Application to revenue claims:
The person who makes qualifying creative contributions may claim copyright ownership and corresponding licensing income in the protected human-authored elements.
Case 8. IceTV Pty Ltd v Nine Network Australia Pty Ltd [2009] HCA 14
Persuasive Australian authority
Facts and legal issue:
The dispute concerned the reproduction of television-program scheduling information.
The High Court considered originality and the relationship between effort expended in gathering information and the expression of a protected work.
Judicial principle:
Copyright does not necessarily reward the expenditure of substantial resources in collecting information. The relevant intellectual contribution must be connected to the form of expression protected by copyright.
Relevance to autonomous creative systems:
An AI developer may spend millions developing a creative model, purchasing computing resources and collecting data.
Those investments do not automatically confer copyright ownership over every output.
Application to revenue claims:
A developer seeking a share of commercial revenue must identify an independent legal basis, such as contractual rights, copyright ownership or another recognised entitlement.
Case 9. Telstra Corporation Ltd v Phone Directories Company Pty Ltd [2010] FCAFC 149
Persuasive Australian authority
Facts and legal issue:
The dispute concerned copyright in telephone directories generated through automated computerised processes.
The court considered whether sufficient human authorship existed in the resulting compilations.
Judicial principle:
Automated production does not automatically satisfy the human authorship and originality requirements.
The court examined the extent to which identifiable human contributions could be attributed to the resulting works.
Relevance to autonomous creative systems:
This decision is especially relevant to systems that generate thousands of images, videos or musical compositions with limited human intervention.
Application to revenue claims:
An AI platform cannot necessarily claim copyright ownership over automatically produced material simply because its technology created the output.
Case 10. Thaler v Perlmutter, 687 F. Supp. 3d 140 (D.D.C. 2023)
Persuasive United States authority
Facts and legal issue:
The dispute concerned an artwork generated autonomously by an artificial intelligence system.
The applicant sought copyright registration identifying the AI system as the author.
Judicial principle:
The United States District Court held that copyright protection under US law requires human authorship.
The decision rejected the proposition that an AI system could independently qualify as the author.
Relevance to autonomous creative systems:
The decision illustrates the legal difficulties surrounding claims that an AI system itself should receive copyright ownership or commercial royalties.
Application to revenue claims:
A Singapore court may find the reasoning informative when considering whether an autonomous system can independently hold copyright, although the US decision is not binding in Singapore.
Case 11. Andy Warhol Foundation for the Visual Arts, Inc. v Goldsmith, 598 U.S. 508 (2023)
Persuasive United States authority
Facts and legal issue:
The dispute concerned the commercial licensing of an artwork derived from a copyrighted photograph.
The US Supreme Court considered whether the particular commercial use was sufficiently transformative for fair-use purposes.
Judicial principle:
The commercial purpose and specific context of the use are important when assessing fair use. A new artistic appearance does not automatically make every commercial use permissible.
Relevance to autonomous creative systems:
An AI-generated artwork may resemble, incorporate or transform existing protected material.
The fact that an AI output looks different from the source material does not automatically resolve infringement questions.
Application to revenue claims:
Original copyright owners may seek compensation where their protected works are unlawfully exploited in commercially distributed AI-generated products.
Case 12. Authors Guild v Google, Inc., 804 F.3d 202 (2d Cir. 2015)
Persuasive United States authority
Facts and legal issue:
The dispute involved Google's digitisation of books and the creation of a searchable digital-book database.
The court considered whether the copying and provision of search functionality constituted fair use under US law.
Judicial principle:
The court found the particular book-digitisation and search functionality to be fair use under the applicable US legal framework.
Relevance to autonomous creative systems:
AI companies may use large collections of books, documents and other materials for computational analysis and training.
This case offers a comparative perspective on the relationship between technological innovation, copying and copyright exceptions.
Application to revenue claims:
It is relevant when evaluating whether the use of copyrighted material in developing a commercial creative platform constitutes infringement. However, US fair-use doctrine cannot simply be imported into Singapore's statutory exception framework.
5. Copyright Revenue Entitlement: Who Can Make a Claim?
The following table identifies the potential claimants and their legal arguments.
| Claimant | Possible legal entitlement |
|---|---|
| Human creator | Copyright ownership and royalties |
| AI developer | Contractual licensing or other established rights |
| Employer | Copyright ownership under employment provisions |
| Commissioning company | Contractual or statutory ownership entitlement |
| Copyright owner of training material | Infringement remedies where infringement is established |
| Exclusive licensee | Statutory infringement action and contractual remedies |
| Distributor | Contractual distribution commissions |
| Music rights holder | Relevant composition, recording or performance royalties |
Important distinction: An AI developer's ownership of software does not automatically give it ownership of every creative output produced using that software.
Similarly, purchasing access to an AI platform does not automatically transfer all intellectual property rights to the customer.
The applicable contract and copyright law must be considered separately.
6. Types of Civil Claims
A. Copyright ownership declaration claims
A claimant may request a judicial declaration determining ownership of a disputed AI-assisted work.
For example, a digital artist may seek a declaration that copyright in a substantially human-created illustration belongs to them rather than the AI platform.
B. Unpaid royalty claims
A creator may bring proceedings where a platform fails to distribute royalties required under an agreement.
Potential claims include:
Outstanding licensing payments.
Underpaid commercial royalties.
Unauthorised deductions.
Breach of revenue-sharing agreements.
Failure to provide contractual accounting statements.
C. Copyright infringement claims
Where protected material is reproduced or commercially exploited without permission, the copyright owner may pursue an infringement action.
Potential remedies include damages, an account of profits where appropriate, injunctions and other statutory remedies.
D. Breach of contract claims
Contracts involving autonomous creative systems should specify who owns the resulting works and who receives commercial income.
A breach may occur where a platform:
Sells content outside the agreed licence.
Fails to pay agreed royalties.
Transfers rights without permission.
Misrepresents ownership.
Uses a creator's content beyond the contractual scope.
E. Misappropriation of confidential creative assets
A company may pursue a confidentiality claim where another party improperly obtains or exploits protected commercial information, such as unpublished creative materials, proprietary prompts or confidential model information.
Such claims are legally distinct from copyright infringement.
7. Evidence Required in Autonomous Creative Copyright Revenue Litigation
Evidence is particularly important because the court may need to reconstruct the creative process and determine which participant contributed the protected expression.
Important evidence includes:
| Evidence | Legal significance |
|---|---|
| Original prompts | Demonstrates human creative instructions |
| AI generation logs | Establishes how the output was produced |
| Source files | Identifies original and modified content |
| Editing history | Demonstrates human creative contribution |
| Employment agreements | Establishes potential ownership |
| AI platform terms | Determines contractual rights |
| Licensing agreements | Establishes royalty entitlements |
| Revenue statements | Calculates unpaid income |
| Training-data records | Helps investigate source-material infringement |
| Distribution records | Identifies commercial exploitation |
Digital evidence may also establish whether the disputed output was generated independently or derived from an existing protected work.
8. Calculation of Copyright Revenue and Damages
The financial value of a claim depends on the nature of the commercial exploitation and the legal basis of the claimant's entitlement.
Common revenue categories include:
Direct sales of AI-generated creative products.
Subscription income attributable to licensed creative content.
Commercial reproduction fees.
Music streaming royalties.
Film and animation licensing fees.
Advertising revenue under applicable contractual arrangements.
Commercial sublicensing payments.
Marketplace distribution income.
For example, consider a hypothetical AI-assisted creative project:
| Revenue category | Amount |
|---|---|
| Digital artwork sales | S$40,000 |
| Commercial licensing | S$30,000 |
| Subscription-related revenue | S$20,000 |
| Advertising revenue | S$10,000 |
| Total revenue | S$100,000 |
If an enforceable agreement provides that the creator receives 40% of the relevant revenue, the creator's contractual entitlement would be S$40,000, subject to the agreement's definitions, permitted deductions and accounting provisions.
This is a contractual illustration, not an automatic statutory royalty rate.
9. Jurisdiction and Enforcement in Singapore
Civil copyright disputes may be brought before the Singapore courts in accordance with the applicable jurisdictional and procedural requirements.
Depending on the claim, the claimant may seek:
A declaration of copyright ownership.
An injunction against unauthorised exploitation.
Damages for copyright infringement.
An account of profits where legally available.
Contractual compensation.
Recovery of unpaid royalties.
Appropriate confidentiality remedies.
Orders relating to disclosure and preservation of evidence.
Section 154 of the Copyright Act 2021 provides a six-year limitation period for copyright infringement actions.
International disputes may create additional issues because AI platforms frequently operate across multiple countries.
The claimant may need to establish the relevant territorial connection, applicable law, jurisdiction and location of the alleged infringing conduct.
10. Regulatory and Institutional Role
Several Singapore institutions are relevant to this area.
A. Intellectual Property Office of Singapore (IPOS)
IPOS provides information and administrative support concerning copyright ownership, licensing, commercialisation and intellectual property policy.
B. Singapore courts
The courts determine copyright ownership disputes, infringement claims, contractual royalty disputes and appropriate civil remedies.
C. Copyright Tribunals
The statutory copyright tribunal framework provides a specialised mechanism for certain copyright licensing disputes within its jurisdiction.
D. Infocomm Media Development Authority (IMDA)
IMDA's work concerning digital and AI governance may be relevant to the broader technological environment in which autonomous creative platforms operate.
These institutions have different responsibilities; none automatically determines the ownership of every AI-generated creative work.
11. Practical Recommendations for Copyright Revenue Protection
Companies and individual creators operating autonomous creative systems should consider the following safeguards.
1. Written ownership agreements
Clearly identify ownership of AI-assisted and AI-generated materials.
2. Human contribution records
Maintain evidence of human creative instructions, selection, editing and arrangement.
3. Transparent royalty arrangements
Specify revenue-sharing percentages, payment schedules and permissible deductions.
4. Training-data compliance
Maintain records of relevant permissions, licences and statutory exceptions.
5. Commercial-use restrictions
Define whether outputs can be used for advertising, publishing, music production, film distribution and sublicensing.
6. Independent financial audits
Establish procedures for verifying platform revenue and royalty calculations.
7. Dispute-resolution provisions
Include suitable mediation, arbitration or litigation clauses.
12. Critical Legal Challenges and Future Development
Singapore's copyright framework faces several emerging questions.
| Emerging issue | Legal challenge |
|---|---|
| Fully autonomous artwork | Identifying a human author |
| AI-generated music | Separating composition and recording rights |
| Autonomous film production | Allocating rights among contributors |
| AI training datasets | Determining lawful use of source material |
| Creative AI marketplaces | Establishing ownership and royalty transparency |
| AI-generated derivatives | Determining substantial reproduction |
| Multiple AI developers | Allocating contractual responsibilities |
| International creative platforms | Cross-border enforcement |
| Autonomous royalty systems | Establishing accountability for incorrect payments |
| AI-created digital assets | Determining copyright and other applicable IP rights |
A particularly important future question is whether Singapore should develop a more specific statutory framework for AI-generated works.
Possible reforms could address:
Attribution of human creative contributions.
Legal treatment of substantially autonomous outputs.
Disclosure of AI involvement.
Rights in AI-assisted derivative works.
Standardised royalty-accounting obligations.
Contractual protection for independent creators.
Licensing arrangements for AI training materials.
These are potential areas for legal development rather than existing general statutory entitlements.
13. Overall Legal Conclusion
Autonomous Creative System Copyright Revenue Claims in Singapore involve the intersection of copyright law, contract law, commercial exploitation, confidentiality and civil liability.
The Copyright Act 2021 provides the principal statutory framework, but traditional copyright principles create important limitations where creative outputs are generated predominantly by autonomous systems.
The most significant judicial authority is Asia Pacific Publishing Pte Ltd v Pioneers & Leaders (Publishers) Pte Ltd, because it addresses the relationship between human authorship and originality.
Other Singapore decisions, including Global Yellow Pages, I-Admin, RecordTV and Virtual Map, provide relevant principles concerning originality, confidential information, technological services and copyright exploitation.
Comparative authorities such as Nova Productions, Telstra, Thaler and CCH Canadian further illustrate the difficulties of allocating copyright ownership and commercial benefits in computer-generated works.
The central legal proposition is that investment in an autonomous creative system does not automatically establish copyright ownership of every output. Copyright revenue entitlement must be established through identifiable legal rights, qualifying authorship, statutory ownership rules, contractual arrangements or another recognised cause of action.
The distinction between owning the AI technology, owning the generated work and receiving contractual revenue from its commercial exploitation will remain fundamental to future Singapore copyright litigation.
Research note: This analysis distinguishes binding Singapore authorities from persuasive foreign decisions. The foreign cases should not be treated as establishing Singapore law, and there is presently no reported Singapore appellate decision directly resolving the complete copyright ownership and royalty entitlement framework for generative AI outputs.

comments