CyberCode.ph · Philippines

Can Copyright Owners Opt Out of AI Training in the Philippines?

Last updated September 28, 2026 · Practical privacy, cybersecurity and technology-law guidance

Last materially reviewed: September 6, 2026

Intellectual Property → AI-Generated Works → AI Training & Enforcement

Direct Answer

There is no general Philippine statute that gives every copyright owner a universal AI-training opt-out right. Rights holders can nevertheless signal non-consent, use contractual restrictions, technical controls, platform-specific settings, licensing notices and ordinary copyright enforcement where protected works are copied without authorization and no exception applies.

Legal Status

Developing. Opt-out mechanisms are becoming important internationally, but Philippine copyright law still turns primarily on ownership, exclusive rights, licensing, exceptions such as fair use, and the facts of the copying.

What an Opt-Out Can Do

An opt-out can create a clear record that a rights holder does not authorize a particular use. Depending on the setting, that may be expressed through website terms, API terms, robots directives, machine-readable signals, account settings, licensing metadata or a direct notice to a developer.

But an opt-out signal is not itself a new copyright law. Its effect depends on the legal and contractual context.

Practical Opt-Out Layers

  • Contract: terms of service can restrict automated collection or model training.
  • Technical controls: authentication, rate limits and crawler controls can limit access.
  • Metadata: creators can attach rights and licensing information to files where supported.
  • Platform tools: some services provide creator controls or AI-training settings.
  • Direct notice: rights holders can send a written reservation of rights or licensing notice.

What an Opt-Out Does Not Guarantee

It does not guarantee that every crawler will obey the signal, that past copies will be deleted, or that a court will treat every technical instruction as independently enforceable. Rights holders should therefore combine technical signals with evidence preservation and a clear licensing/enforcement strategy.

Recommended Rights-Holder Workflow

  1. Document ownership and publication history.
  2. Publish a clear AI-training policy or reservation of rights.
  3. Use available technical controls without breaking legitimate indexing needs.
  4. Keep server logs where feasible.
  5. Identify likely model or dataset uses.
  6. Send targeted notices rather than broad unsupported accusations.
  7. Escalate through platform, contractual or IPOPHL enforcement routes where appropriate.

SEO Warning: Do Not Block Search by Accident

Website owners should separate ordinary search-engine indexing from AI-training preferences where tools permit. A blanket block can reduce discoverability. The legal and technical layers should be designed together rather than treating robots.txt as a substitute for copyright strategy.

Related Guides

Choose the Control That Matches Your Goal

Start by deciding whether you want to restrict future collection, stop use within a service, request deletion of retained copies, negotiate a license or challenge alleged copying. These are different objectives. Copyright economic rights and exceptions remain the legal framework; a preference signal alone does not resolve them. IP Code, Sections 177 and 185.

Goal Possible measure Limit
State a reservation Clear policy or licensing notice Does not establish every element of a legal claim
Influence compliant crawling Appropriate crawler directives Not every crawler complies
Restrict access Authentication and access controls Does not recall previously obtained copies
Use a service-specific option Provider’s documented setting Scope depends on the service and current terms
Seek correction or deletion Targeted documented request Entitlement and available response depend on facts and law

The Robots Exclusion Protocol, RFC 9309 is not an access-authorization mechanism. A robots entry should not be relied on to keep confidential material private. Equally, absence of an entry should not be presented as a universal grant of copyright permission.

Three Hypothetical Rights-Holder Situations

A Publisher Wants Search Visibility but Restricts Training

Inventory the controls actually supported by relevant services and their documented scope. Test changes before broad deployment. Avoid a blanket rule that accidentally prevents desired discovery. A provider may distinguish uses differently from another provider; do not copy settings from an unrelated platform and assume identical behavior.

An Illustrator Suspects Past Collection

A new notice can document a present position, but does not establish when earlier collection occurred or automatically alter earlier permissions. Preserve original files, publication history and any evidence of access. Ask the identified recipient to explain the relevant dataset or use without asserting unproven facts.

A Client Upload Was Used Under a Provider Agreement

The client should inspect the terms and settings applicable at upload, who authorized the upload and what the service promises about future use. A setting changed today may not have the same effect on past copies. Seek written clarification rather than promising that disabling a toggle removes every past influence on a model.

What to Include in a Targeted Request

  1. Identify the works with URLs, filenames or other precise identifiers.
  2. Explain your authority as owner or authorized representative.
  3. Describe the evidence and distinguish suspicion from confirmed use.
  4. State the requested action: investigate, stop future collection, exclude data, delete copies or discuss licensing.
  5. Identify the applicable agreement or rights basis where known.
  6. Ask for confirmation of scope, response and unresolved limitations.
  7. Retain the sent request, receipt and replies.

This is a request checklist, not an automatic statutory demand. Do not insert an invented legal response deadline or threaten a fixed penalty without a verified basis. If personal data is involved, privacy rights may require a separate analysis; copyright ownership and data-subject status are not interchangeable.

Keep Evidence of the Policy and Its Limits

Retain dated copies of notices, settings, relevant terms and deployment records. Keep appropriately protected logs where available. Record what each control was intended to cover. A screenshot of a toggle proves a displayed setting at that time; it may not establish the underlying provider’s complete data practices.

For technical investigation, avoid unauthorized access or collecting unnecessary third-party information. Use the evidence guide to organize suspected copying and the scraping guide to distinguish access records from proof of training.

When the Recipient Does Not Agree

Identify whether the dispute concerns ownership, license scope, an exception, technical feasibility or evidence of use. These require different responses. A developer may dispute your claim without that disagreement proving infringement. Preserve the correspondence and seek qualified assessment before making public accusations or filing a formal claim.

A negotiated training license, narrower use or voluntary exclusion may be possible. None should be described as guaranteed. Keep any resolution precise about future collection, stored copies and model-related measures so the parties do not agree to different understandings of “removal.”

Frequently Asked Questions

Does robots.txt create copyright protection?

No. Copyright exists independently. Robots instructions are technical signals and can be evidence of access preferences, but they do not replace the Copyright Code.

Can I demand deletion from a training dataset?

You can request it. Whether the developer must comply depends on the legal basis, contract, platform policy, privacy considerations if personal data is involved, and other facts.

Should creators add “No AI Training” to their terms?

It can help clarify intent and contractual expectations, but wording, assent, enforceability and technical implementation matter.

Official Sources

Important: This article provides general educational information about Philippine law and technology. It is not legal advice and does not create an attorney-client relationship. Laws, agency procedures, platform terms and the facts of each situation may change the result. Verify current requirements through the cited official sources and seek qualified professional advice when your rights, deadlines, money or legal exposure may be affected.

Featured image: Photo by Steve A Johnson via Unsplash.

CyberCode updates

Get practical updates on Philippine technology law, data privacy, cybersecurity, and AI.

Email activity tracking

Unsubscribe any time. See our privacy policy below.