HR professionals reviewing an AI-generated performance report together in a Philippine office, illustrating human review of algorithmic HR tools

Can an Algorithm Legally Fire You? AI in Hiring, Scoring, and Termination Under PH Law

Yes — but only within the same legal framework that governs a decision made by a human manager. Philippine law does not treat AI scoring, screening, or termination-flagging tools as a separate legal category. If an algorithm’s output leads to hiring rejection or dismissal, the employer must still show a valid just or authorized cause under the Labor Code and follow the twin-notice due process rule.[1] No Philippine court has yet ruled on a case squarely involving algorithmic termination, so this guide applies existing, settled doctrine to that new fact pattern rather than citing a case that does not exist. Separately, the National Privacy Commission (NPC) requires employers using AI tools to process personal data lawfully, transparently, and with a working human-intervention mechanism.[6][7]

Direct Answer

An employer may lawfully use AI tools — resume screening, interview scoring, productivity monitoring, or performance-flagging software — to inform hiring and termination decisions, provided two independent legal tracks are satisfied. First, under labor law, the outcome must still rest on a just cause (Labor Code, Art. 297) or authorized cause (Art. 298–299), and the employer must still give the two required written notices and an opportunity to be heard before dismissal — this is the twin-notice rule under Article 292(b) of the Labor Code.[1] The Supreme Court has held that even a substantively valid dismissal is legally defective, and exposes the employer to nominal-damages liability, if the notice-and-hearing procedure is not followed — regardless of how the decision to terminate was reached.[2]

Second, under privacy law, because AI hiring and performance tools necessarily process personal data, the employer is a personal information controller under the Data Privacy Act of 2012 and must have a lawful basis for that processing (commonly “legitimate interest”), apply data minimization and proportionality, disclose the AI system’s use and purpose to affected workers, and maintain a real mechanism for human review and for the worker to object.[6][7] An AI tool can flag, score, or recommend — the employer, not the algorithm, remains the party legally accountable for the hiring or termination decision and for any resulting Labor Code or Data Privacy Act violation.

Decision Snapshot

Question Short answer
Who does this apply to? All private-sector employers and employees in the Philippines using or subject to AI-assisted hiring, scoring, monitoring, or termination-recommendation tools.
Does AI involvement change the legal test for a valid dismissal? No. The employer must still prove just cause or authorized cause and follow the twin-notice procedure, regardless of whether a human or a system generated the recommendation.
Key deadline or requirement First written notice must give the employee a reasonable period (commonly treated as at least 5 calendar days) to submit a written explanation before any hearing or second notice.[3]
Principal exception Authorized-cause terminations (e.g., redundancy from automation) use a different notice timeline — at least 30 days’ written notice to both the employee and DOLE — and require separation pay.[1]
Evidence to retain The AI tool’s output/report, the underlying scoring criteria if disclosed, the notice-to-explain, the employee’s written explanation, minutes of any hearing, the notice of decision, and any NPC-required privacy notice or data-processing disclosure.
First action Request a written explanation of the specific cause relied upon and the evidence (including any AI-generated report) supporting it before responding or resigning.

Key Takeaways

  • Philippine labor law does not create a separate legal standard for AI-assisted decisions — the same just-cause/authorized-cause and twin-notice due-process rules apply.[1]
  • A dismissal can be valid in substance (real cause exists) but still expose the employer to nominal damages if the notice-and-hearing procedure was skipped or rushed, whether or not the decision originated from software.[2]
  • Because AI hiring/HR tools process personal data, employers are separately bound by the Data Privacy Act of 2012 and NPC guidance requiring transparency, proportionality, and a human-intervention/contestation mechanism.[6][7]
  • The NPC has already evaluated one real Philippine case — AI-based scoring of BPO call-center employee performance — and found it permissible under the “legitimate interest” basis, subject to necessity, proportionality, and the employee’s right to object.[6]
  • No Philippine court has yet decided a case specifically about an algorithm-driven termination. This article applies existing doctrine to that fact pattern; it does not cite a case that does not exist.
  • Employees who believe an AI tool contributed to an unfair outcome can request the underlying explanation for the decision, object to automated processing under the Data Privacy Act, and pursue the standard labor remedies (SEnA, NLRC) if dismissed.
Share this guide
Facebook LinkedIn WhatsApp

Jump to a Section

Authority Classification Rule supported Binding effect Official source
Labor Code of the Philippines, Art. 297 (formerly Art. 282) Labor Code provision Just causes for termination (serious misconduct, willful disobedience, gross neglect, fraud/breach of trust, commission of a crime, analogous causes) Binding statute Presidential Decree No. 442 (Labor Code), Official Gazette / LawPhil text
Labor Code of the Philippines, Art. 298–299 (formerly Art. 283–284) Labor Code provision Authorized causes (installation of labor-saving devices, redundancy, retrenchment, closure, disease) and related separation-pay and 30-day dual-notice requirements Binding statute Presidential Decree No. 442 (Labor Code), Official Gazette / LawPhil text
Labor Code of the Philippines, Art. 292(b) (formerly Art. 277(b)) Labor Code provision Procedural due process requirement — written notice and opportunity to be heard before termination for just cause Binding statute Presidential Decree No. 442 (Labor Code), Official Gazette / LawPhil text
Agabon v. NLRC, G.R. No. 158693 Jurisprudential doctrine A dismissal with a valid substantive cause but without proper procedural due process remains a valid dismissal, but the employer is liable for nominal damages for the due-process violation Controlling Supreme Court precedent (En Banc) LawPhil, G.R. No. 158693, Nov. 17, 2004
King of Kings Transport, Inc. v. Mamac, G.R. No. 166208 Jurisprudential doctrine Sets out the detailed content requirements of the twin-notice rule (first notice contents, minimum period to respond, hearing/conference, second notice of decision) Controlling Supreme Court precedent LawPhil, G.R. No. 166208, June 29, 2007
Republic Act No. 10173, Data Privacy Act of 2012 Statute General data-protection obligations of employers as personal information controllers, including lawful basis for processing (e.g., legitimate interest) and data-subject rights Binding statute National Privacy Commission, Data Privacy Act of 2012
NPC Advisory Opinion No. 2024-005 (May 21, 2024) Agency guidance AI-based analysis/scoring of BPO employee calls and emails for performance evaluation can proceed on a “legitimate interest” basis if necessary, proportionate, and subject to the employee’s right to object Advisory opinion — persuasive/interpretive, addressed to the specific requesting party, not a generally binding rule NPC Advisory Opinion No. 2024-005 (PDF)
NPC Advisory No. 2024-04, “Guidelines on the Application of the DPA… to Artificial Intelligence Systems” (Dec. 19, 2024) Administrative rule / agency guidance Transparency, accountability, fairness (anti-“AI washing”), data minimization, mandatory human-intervention mechanism, and data-subject rights to object/rectify/erase for AI systems processing personal data NPC advisory guidance — interpretive, not a statute NPC Advisory No. 2024-04 (PDF)

Constitutional basis

Article XIII, Section 3 of the 1987 Constitution directs the State to afford full protection to labor and to guarantee security of tenure. This constitutional policy underlies both the just-cause/authorized-cause requirement and the due-process protections discussed below; it establishes state policy and is implemented through the Labor Code and jurisprudence rather than creating a directly enforceable technology-specific rule.

Labor Code basis

Under current Labor Code numbering (post-2015 renumbering, discussed below), the relevant termination provisions are Article 297 (just causes), Articles 298–299 (authorized causes), and Article 292(b) (due process). None of these provisions distinguishes between a human-made and an algorithm-assisted decision. The Supreme Court has confirmed the article renumbering; for example, one 2024 decision expressly states that “Article 292 (formerly Article 277) of the Labor Code provides for the manner of termination of employment based on just cause.”[3]

Relevant statutes

The Data Privacy Act of 2012 (RA 10173) governs any employer’s collection, use, and processing of employee personal data through AI hiring or monitoring tools, independently of whether the underlying employment action is also valid under the Labor Code. A hiring or termination decision can be procedurally compliant under labor law and still expose the employer to a separate Data Privacy Act violation (or vice versa) — the two frameworks are cumulative, not substitutes for each other.

Administrative issuances

NPC Advisory Opinion No. 2024-005 and NPC Advisory No. 2024-04 are the two primary agency issuances currently addressing AI use that intersects with employment.[6][7] Both are NPC guidance, not statutes — an advisory opinion is addressed to the specific facts presented by the requesting party and is persuasive rather than automatically binding on other employers with different facts; the AI Guidelines are a general interpretive framework applying existing DPA obligations to AI systems. As of this article’s publication date, no DOLE department order or labor advisory specifically governing AI use in hiring or termination decisions could be confirmed as issued. DOLE officials have publicly referenced work on this subject, but this guide does not cite a specific numbered DOLE AI advisory because none could be verified against an official DOLE source — treat this as an open regulatory development to watch, not a settled rule.

Jurisprudence

Agabon v. NLRC, G.R. No. 158693, decided En Banc on November 17, 2004: the employees were dismissed for abandonment (a just cause), but the employer failed to serve the required written notices. The Supreme Court upheld the dismissal as valid in substance — abandonment was proven — but held that the failure to follow the twin-notice procedure violated the employees’ statutory due-process rights. Instead of ordering reinstatement and backwages, the Court awarded nominal damages (₱30,000 per employee in that case) to vindicate the procedural violation.[2] The doctrinal significance for AI-assisted decisions: a valid cause for termination — even one flagged or substantiated by an algorithmic tool — does not excuse the employer from separately completing the notice-and-hearing procedure. Skipping that procedure because “the system already decided” would expose the employer to the same nominal-damages liability described in Agabon.

King of Kings Transport, Inc. v. Mamac, G.R. No. 166208, decided June 29, 2007: the Supreme Court detailed what the twin-notice rule actually requires in practice — a first written notice specifying the particular acts or omissions charged, giving the employee a reasonable opportunity (generally treated as at least five calendar days) to submit a written explanation; a hearing or conference where circumstances warrant; and a second written notice stating the employer’s findings and, if applicable, the grounds for termination.[3] These content requirements apply identically whether the underlying cause was identified by a supervisor’s personal observation or by an AI performance-monitoring report.

Applicability and Decision Path

Use the following questions to assess whether an AI-assisted hiring or termination action is legally sound. This is general information to help identify what facts matter — it is not a determination of any specific case.

Question 1: Is there an employer-employee relationship (or a prospective one, for hiring)?
Question 2: What specific just cause (Art. 297) or authorized cause (Art. 298-299) is being relied upon?
Question 3: Did the AI tool merely inform/flag the decision, or did the employer treat its output as automatically final with no human review?
Question 4: Were the twin-notice steps (first notice, opportunity to explain, hearing where warranted, second notice) actually completed, and on what timeline?
Question 5: Was the worker told, in substance, that an AI/automated tool was used to generate the score, flag, or recommendation relied upon?
Question 6: Did the worker have a real opportunity to contest or seek human review of the AI-generated output, consistent with the NPC's human-intervention requirement?
Question 7: For hiring, was the rejection based on a protected characteristic the AI tool may have used as a proxy variable (e.g., an address correlating with a protected class), which could raise separate discrimination concerns?

Coverage generally extends to all private-sector rank-and-file and supervisory employees under the Labor Code, including probationary employees (subject to their own separate due-process standard) and BPO/contact-center staff whose calls or output are commonly subject to AI-based quality scoring.[6] Managerial employees have a modified due-process framework under separate jurisprudence and are not the focus of this guide. Excluded from Labor Code coverage in the ordinary sense are genuine independent contractors and workers outside an employer-employee relationship, though platform/gig-worker classification is itself a contested, fact-specific question addressed in a separate LaborCode.ph guide.

Doctrinal Analysis

Twin-Notice Rule

Definition

The twin-notice rule is the procedural due-process standard requiring an employer to serve two separate written notices — a notice to explain (first notice) and a notice of decision (second notice) — with an intervening opportunity to be heard, before terminating an employee for just cause.

Source

Labor Code, Art. 292(b) (formerly Art. 277(b)); elaborated in King of Kings Transport, Inc. v. Mamac, G.R. No. 166208 (2007).[3]

Elements

  • A first written notice specifying the particular acts or omissions and the specific cause for termination being considered
  • A reasonable period (commonly treated as at least five calendar days) for the employee to submit a written explanation
  • A hearing or conference where the circumstances require it
  • A second written notice stating the employer’s findings and the decision

Legal test

Courts assess whether both notices were actually served, whether the first notice gave adequate factual detail (not a vague or generic accusation), and whether the employee had a genuine, not merely nominal, chance to respond before the decision was finalized.

Burden of proof

The employer bears the burden of proving both substantive cause and compliance with procedural due process; failure to prove either exposes the employer to liability, though the specific remedy differs (illegal dismissal remedies for lack of cause; nominal damages for a procedural lapse where cause is otherwise proven).[2]

Exceptions

Authorized-cause terminations (redundancy, retrenchment, closure, disease, labor-saving devices) use a different notice framework: at least 30 days’ written notice to both the affected employee and the DOLE regional office, plus separation pay, rather than the just-cause twin-notice sequence.[1]

Defenses

An employer may argue substantial compliance where notices were delayed by circumstances beyond its control, but courts generally construe the notice requirements strictly given their constitutional due-process foundation.

Remedies

Nominal damages where cause is valid but procedure was defective (Agabon); reinstatement, backwages, or separation pay in lieu of reinstatement where no valid cause existed at all.

Related doctrines

Just cause, authorized cause, illegal dismissal, and (for hiring-stage issues) the general prohibition on discriminatory hiring practices.

Factual patterns

Employers relying on system-generated productivity or quality-assurance reports sometimes treat the report itself as the “decision,” skipping a genuine notice-and-hearing sequence. Nothing in Philippine jurisprudence supports treating an AI report as a substitute for the notice-and-hearing steps; the cases above evaluate whether the human-run process was followed, not what generated the underlying evidence.

AI Use in HR Decisions Under the Data Privacy Act

Definition

Where an employer uses software that processes employee or applicant personal data (call recordings, keystroke logs, performance metrics, resumes, interview video) to generate a score, ranking, or recommendation, the employer is a personal information controller under RA 10173 and must independently satisfy the DPA’s processing, transparency, and data-subject-rights requirements.

Source

RA 10173; NPC Advisory Opinion No. 2024-005 (May 21, 2024); NPC Advisory No. 2024-04 (Dec. 19, 2024).[6][7]

Elements

  • A lawful basis for processing (in the reviewed BPO scoring case, “legitimate interest” under the DPA’s implementing rules)
  • Necessity and proportionality — data collected must be adequate, relevant, and not excessive relative to the stated purpose
  • Transparency — employees informed of the AI system’s purpose, nature, and extent of processing in plain language
  • A working human-intervention and contestation mechanism for AI-assisted decisions with significant effects on the worker
  • A functioning channel for the data subject to object, seek rectification, or request erasure

Legal test

The NPC’s May 2024 advisory opinion did not issue a blanket rule that all AI performance-scoring is automatically lawful; it held that legitimate interest was a viable basis on the specific facts presented, while expressly stating that weighing the employer’s legitimate interest against a specific employee’s objection is a determination the employer (as personal information controller) must make, not something the NPC decides in the abstract for every case.[6]

Burden of proof

The employer, as personal information controller, bears the burden of demonstrating a lawful basis, necessity, and proportionality if a data-subject complaint is filed with the NPC.

Exceptions

The Dec. 2024 AI Guidelines advisory is a general cross-sector framework; it does not contain provisions written specifically for employment or HR use, so its human-intervention and transparency requirements must be applied to the HR context by extension rather than through an employment-specific NPC rule.[7]

Defenses

Documented necessity, proportionality analysis, and a demonstrable human-review step are the strongest available defenses to an NPC complaint concerning AI-assisted HR processing.

Remedies

NPC complaint and investigation process; potential administrative and criminal penalties under RA 10173 for data-privacy violations, separate from any labor-law remedy.

Related doctrines

Data subject rights (right to object, right to be informed, right to access), personal information controller obligations, twin-notice rule (labor-law track runs in parallel, not as a substitute).

Factual patterns

The one Philippine AI-in-HR case the NPC has actually evaluated involved a BPO subsidiary of a foreign hospitality company using AI to analyze call-center employee calls and emails for performance scoring — a fact pattern very close to common Philippine contact-center quality-assurance practice.[6] No comparable NPC opinion yet addresses AI-assisted hiring screening specifically; the transparency and human-intervention principles from the Dec. 2024 Guidelines are the best available primary-source basis for analyzing that scenario, applied by extension.

Evidence and Documentation

Record Who typically controls it Why it matters
AI tool’s output report, score, or flag that led to the action Employer / HR / vendor Establishes what the tool actually said and whether it was treated as advisory or automatically final
Written policy or notice disclosing use of an AI tool in hiring, scoring, or monitoring Employer Supports the DPA transparency requirement; its absence is a weakness for the employer in an NPC complaint
Notice to explain (first notice) and employee’s written response Both parties (employer serves; employee should keep a copy) Core evidence of twin-notice compliance and of the substantive cause alleged
Minutes or record of any hearing/conference Employer Shows the employee had a genuine opportunity to be heard, not just a formality
Notice of decision (second notice) Employer Establishes when and on what stated grounds the employer finalized the action
Employment contract, job posting, or offer with any AI-screening disclosure Both parties Relevant to hiring-stage disputes over automated rejection
Payroll, timekeeping, and performance-metric history predating the AI flag Employer (employee should request copies) Helps assess whether the AI-flagged issue is consistent with, or contradicted by, the worker’s actual record
Any objection or data-subject request the employee submitted regarding the AI processing Employee (keep proof of submission/receipt) Preserves the DPA right-to-object claim and shows the employer’s response, or lack of one

Employees should preserve their own copies of notices, correspondence, and any information disclosed about the AI system used; employers should retain the underlying AI output alongside the human review record, since a bare AI score without a documented human decision step weakens both the labor-law and privacy-law defense. Neither party should alter, delete, or fabricate a record — doing so can independently undermine credibility in a SEnA conciliation, NLRC case, or NPC complaint regardless of the merits of the underlying dispute.

Calculation and Deadline Support

Not applicable in the ordinary sense for this topic. This guide does not involve a wage, benefit, or contribution calculation. The only deadline-type element is procedural: the reasonable period (commonly treated as at least five calendar days) for an employee to respond to a first notice under the twin-notice rule, and the at-least-30-day dual notice (to the employee and to DOLE) required for authorized-cause terminations such as redundancy.[1][3] Where an authorized-cause termination results from automation or an AI-driven efficiency initiative, the separation-pay formula is the same as for any other authorized-cause termination (redundancy: at least one month’s pay or one month’s pay per year of service, whichever is higher; retrenchment or closure not due to serious losses: at least one-half month’s pay per year of service, whichever is higher) — that computation is addressed in LaborCode.ph’s dedicated separation-pay guide rather than repeated here, since introducing a full worked example would duplicate that canonical resource rather than serve this article’s distinct intent.

Document and Communication Support

Document Purpose Who prepares it When used
Notice to explain (first notice) Formally identifies the specific act, omission, or performance issue — including one flagged by an AI system — and starts the response period Employer/HR Before any just-cause termination
Employee’s written explanation Employee’s formal response to the notice to explain; may request the underlying AI report or scoring methodology Employee Within the period stated in the first notice
Request for AI-processing information / data-subject request Invokes the DPA right to be informed and, where applicable, the right to object to automated processing Employee On learning that an AI tool contributed to a decision affecting them
Notice of decision (second notice) States the employer’s findings and final action Employer/HR After the hearing/explanation period
30-day authorized-cause notice (to employee and DOLE) Required for redundancy/retrenchment/automation-driven separations Employer/HR At least 30 days before the effective date of separation

None of these documents guarantees a particular legal outcome; their purpose is to create a contemporaneous, accurate record that either party can rely on if the matter proceeds to SEnA, the NLRC, or an NPC complaint. Aggressive or inaccurate wording in a notice or response can itself become evidence used against the drafting party.

What to Do Next

  1. Preserve every notice, report, and piece of correspondence related to the decision, including any AI-generated score or report if it was shared or referenced.
  2. Request, in writing, the specific cause relied upon and whether an AI or automated tool contributed to the decision — this supports both a labor-law due-process check and a Data Privacy Act transparency request.
  3. Compare what actually happened against the twin-notice elements above (first notice with specifics, reasonable response period, hearing where warranted, second notice) or, for authorized-cause cases, the 30-day dual-notice and separation-pay requirements.
  4. If a data-privacy concern exists — no disclosure of AI use, no way to object, no human review available — consider a written objection to the employer as personal information controller before escalating externally.
  5. Attempt an internal correction or grievance process where one exists, or use the Single Entry Approach (SEnA) at the DOLE for a mandatory 30-day conciliation-mediation before formal litigation.
  6. If unresolved, evaluate filing an illegal dismissal complaint with the NLRC (labor-law track) and/or a complaint with the National Privacy Commission (data-privacy track) — the two tracks are independent and can proceed in parallel.
  7. Consult a Philippine labor lawyer when the facts are contested, the amount at stake is significant, or the AI-related evidence is technical or disputed.

Practical Hypotheticals

Hypothetical 1: AI-scored call quality leads to termination at a BPO

Facts: A call-center agent’s calls are analyzed by an AI quality-assurance tool that flags a pattern of “low customer sentiment scores” over two months. HR issues a termination letter citing “poor performance per QA system flag,” with no prior notice to explain and no hearing.

Legal issue: Whether the termination satisfies both substantive cause (if performance-based, this is typically argued as gross and habitual neglect of duties or an analogous cause under Art. 297) and procedural due process.

Applicable rule: Art. 297 (just cause) and Art. 292(b) (twin-notice rule); Agabon and King of Kings Transport.[1][2][3]

Evidence that matters: The AI QA reports themselves, any prior coaching or performance-improvement documentation, and — critically — proof (or absence) of a notice to explain and hearing.

Likely analysis: Even if the AI-flagged performance issue is real and well-documented, skipping the notice-and-hearing sequence is a procedural due-process defect under Agabon-line doctrine, exposing the employer to nominal-damages liability at minimum; if the underlying cause itself is not well-substantiated, the case may instead be analyzed as illegal dismissal with reinstatement/backwages exposure.

Important missing facts: Whether the employee received any prior notice of the QA standard, whether the AI scoring methodology was disclosed, and whether the agent had any opportunity to contest the score before termination.

Document or communication needed: Notice to explain referencing the specific QA data; employee’s written response; hearing record; second notice.

Possible next step: The employee can request the underlying QA data and file a SEnA request; the employer should audit whether its AI-QA termination process actually includes a documented notice-and-hearing step before relying on it again.

Hypothetical 2: AI applicant-tracking system auto-rejects a job application

Facts: An applicant is auto-rejected by an AI resume-screening tool before any human reviewer sees the application. The applicant later learns, through a LinkedIn post by the company, that the tool screens out candidates with employment gaps longer than six months.

Legal issue: Whether this raises a Data Privacy Act transparency/human-intervention issue, and whether the screening criterion could function as a proxy for a protected characteristic (e.g., disability, pregnancy, or caregiving-related gaps).

Applicable rule: Because there is no employer-employee relationship yet, the Labor Code’s twin-notice and just-cause rules do not apply to a hiring rejection; the relevant frameworks are the Data Privacy Act (if personal data was processed without adequate transparency or human-review mechanisms) and general anti-discrimination principles, which are addressed in other LaborCode.ph guides rather than this one.[6][7]

Evidence that matters: The employer’s public statements about its screening criteria, any privacy notice provided at application, and whether a human reviewer was ever involved.

Likely analysis: This is an open area without a specific NPC opinion on AI hiring-screening; the Dec. 2024 AI Guidelines’ transparency and human-intervention principles are the best available primary-source analogy, but no confirmed NPC ruling on this exact hiring scenario exists as of this article’s publication.

Important missing facts: Whether the applicant was told an AI tool would screen the application, and whether a review or appeal mechanism existed.

Document or communication needed: A written request to the employer, as personal information controller, for information about the automated processing used.

Possible next step: The applicant may raise a data-subject request/complaint with the NPC; this scenario does not support a labor-law dismissal claim because no employment relationship existed.

Hypothetical 3: Redundancy after adopting an AI-driven workflow

Facts: A finance company adopts an AI document-processing tool that eliminates the need for three data-entry positions. The company issues redundancy notices citing the new software, with 45 days’ advance notice to the affected employees and to DOLE, and offers separation pay computed at one month’s pay per year of service.

Legal issue: Whether automation-driven redundancy is a valid authorized cause and whether the notice and separation-pay requirements were met.

Applicable rule: Art. 298 (formerly Art. 283) — installation of labor-saving devices/redundancy as authorized causes, with the accompanying 30-day dual-notice and separation-pay requirements.[1]

Evidence that matters: A good-faith redundancy study or business justification, fair and reasonable selection criteria for which positions were affected, proof of the 30-day notices to both the employees and DOLE, and proof of separation-pay computation and payment.

Likely analysis: On these facts, the redundancy appears procedurally compliant (notice period exceeds the 30-day minimum; separation pay matches the redundancy formula). The employer still needs a documented, good-faith business rationale — “we adopted AI” is not automatically sufficient without evidence the redundancy was not used as a pretext for a different, improper reason.

Important missing facts: The selection criteria used to choose which employees were declared redundant, and whether comparable positions were created elsewhere in the company around the same time (which could undercut the good-faith claim).

Document or communication needed: The formal 30-day notices, the redundancy study/justification, and the separation-pay computation sheet.

Possible next step: Affected employees who believe the redundancy was not genuine (e.g., positions were simply refilled under a different title) may consult a labor lawyer about challenging the authorized cause itself, separate from any procedural compliance question.

Fact Tables

Aspect Just Cause (Art. 297) Authorized Cause (Art. 298–299)
Basis Employee’s own act or omission (misconduct, neglect, breach of trust, etc.) Business or health reasons not attributable to employee fault (redundancy, retrenchment, closure, disease, automation)
Notice requirement Twin-notice: first notice, opportunity to explain/hearing, second notice 30-day written notice to both employee and DOLE regional office
Separation pay Generally none required by law (though company policy may provide it) Required — amount depends on the specific authorized cause
Does AI involvement change the rule? No — same substantive and procedural standard applies No — same substantive and procedural standard applies

Terminology

Term Plain-English meaning Legal classification
Twin-Notice Rule The requirement to give two written notices, with a chance to be heard in between, before firing someone for just cause Procedural rule (Labor Code, jurisprudence) — see the Twin-Notice Rule glossary entry
Just Cause A termination reason based on the employee’s own fault or misconduct Labor Code provision — see the Just Cause glossary entry
Authorized Cause A termination reason based on business or health circumstances, not employee fault Labor Code provision — see the Authorized Cause glossary entry
Personal Information Controller The party (here, the employer) that decides how and why personal data is processed Data Privacy Act term
Legitimate Interest A lawful basis for processing personal data where the controller’s genuine business interest is not outweighed by the data subject’s rights Data Privacy Act / NPC Circular 2023-07 concept
Human-Intervention Mechanism A real, working process letting a human review or override an AI-generated decision NPC Advisory No. 2024-04 requirement
Algorithmic Management Using software/AI to direct, monitor, evaluate, or discipline workers, in whole or in part Editorial/descriptive term — not yet a defined statutory term in Philippine law

Frequently Asked Questions

Can an employer legally fire someone based solely on an AI system’s recommendation?

An AI system’s output can inform the decision, but the employer must still independently establish a valid just or authorized cause and complete the twin-notice due-process steps. Treating the AI output as the final decision, with no human review and no notice-and-hearing process, does not satisfy Philippine labor law.[1][2]

Is there a Philippine law that specifically regulates AI in hiring or firing?

Not yet, as a dedicated statute. The applicable framework is the existing Labor Code (just cause, authorized cause, twin-notice due process) plus the Data Privacy Act and NPC guidance (Advisory Opinion No. 2024-005 and Advisory No. 2024-04) to the extent personal data is processed by the AI tool.[1][6][7]

Has any Philippine court ruled on an algorithm-driven termination?

No confirmed Philippine court decision squarely addresses algorithmic or AI-driven termination as of this article’s publication. This guide applies existing due-process and just-cause/authorized-cause doctrine to that fact pattern; it does not rely on a case that does not exist.

Does the NPC require employers to disclose that they use AI in HR decisions?

NPC Advisory No. 2024-04 requires transparency about the nature, purpose, and extent of AI processing, and requires a working human-intervention mechanism for AI-assisted decisions with significant effects on individuals — principles that apply to HR use even though the advisory has no employment-specific chapter.[7]

What is the “legitimate interest” basis the NPC discussed for AI performance scoring?

In NPC Advisory Opinion No. 2024-005, the NPC found that a company’s use of AI to analyze and score call-center employee calls/emails for performance evaluation could rely on “legitimate interest” as a lawful basis for processing, provided the processing was necessary, proportionate, adequately disclosed, and subject to the employee’s right to object — but the NPC noted that weighing the employer’s interest against a specific employee’s objection is ultimately for the employer (as controller) to determine on the facts, not a blanket approval for every AI-scoring scenario.[6]

Can an employee object to being scored or monitored by an AI tool at work?

Yes — the Data Privacy Act and NPC guidance recognize a data subject’s right to object to processing, including AI-assisted processing, though the employer may be able to continue processing if another lawful ground applies (such as necessity for the employment relationship). This is a case-by-case determination, not an absolute veto.[6][7]

Does using AI change the redundancy/separation-pay rules if a job is automated away?

No. Automation-driven redundancy still follows the ordinary authorized-cause framework: a good-faith business rationale, fair selection criteria, at least 30 days’ written notice to the employee and DOLE, and the applicable separation-pay formula under Art. 298–299.[1]

What should an employee do first if they suspect an AI tool caused an unfair outcome?

Request, in writing, the specific cause relied upon and whether an AI or automated system contributed to the decision. This supports both a labor-law due-process review and a Data Privacy Act transparency request, and creates a documented starting point before pursuing SEnA, the NLRC, or an NPC complaint.

Is DOLE expected to issue AI-specific labor rules?

DOLE officials have publicly referenced work on AI-related labor guidance, but no finalized, numbered DOLE department order or labor advisory specifically governing AI use in hiring or termination could be confirmed as issued as of this article’s publication. This is an open regulatory development to monitor, not a citable rule.

Conclusion

An algorithm can inform, score, flag, or recommend — but it cannot substitute for the legal requirements Philippine law places on the employer. Every AI-assisted hiring or termination decision must still rest on a genuine just cause or authorized cause under the Labor Code, and every just-cause dismissal still requires the twin-notice procedure regardless of what generated the underlying evidence.[1][2][3] Because AI hiring and monitoring tools process personal data, employers face a second, independent layer of obligations under the Data Privacy Act and NPC guidance — transparency, proportionality, and a working human-intervention mechanism.[6][7] No Philippine court has yet decided a case squarely on algorithmic termination, and no dedicated DOLE AI-labor rule has been confirmed as issued; both are genuine open questions worth monitoring rather than settled law to cite. Employees and employers alike should document the human review step behind any AI-assisted employment decision — that documentation is what will matter if the decision is ever challenged.

Constitution and statutes

[4] 1987 Constitution of the Philippines, Article XIII, Section 3 (protection to labor, security of tenure), Official Gazette, https://www.officialgazette.gov.ph/constitutions/1987-constitution/. Supports: constitutional policy basis for security of tenure and due process. Status: verified official source.

[5] Republic Act No. 10173, Data Privacy Act of 2012, National Privacy Commission, https://privacy.gov.ph/data-privacy-act/. Supports: employer obligations as personal information controller when processing employee/applicant data via AI tools. Status: verified official source.

Labor Code provisions

[1] Labor Code of the Philippines, Presidential Decree No. 442, Articles 297, 298, 299, and 292(b) (current numbering; formerly Articles 282, 283, 284, and 277(b) respectively, per DOLE Department Advisory No. 01, Series of 2015 renumbering), LawPhil Philippine Laws and Jurisprudence Databank, https://lawphil.net/statutes/presdecs/pd1974/pd_442_1974.html. Supports: just-cause and authorized-cause grounds for termination, and the twin-notice/30-day notice due-process requirements. Status: verified official source; article renumbering independently confirmed via Supreme Court decision citing “Article 292 (formerly Article 277).”

Supreme Court decisions

[2] Jenny M. Agabon and Virgilio C. Agabon v. National Labor Relations Commission, Riviera Home Improvements, Inc., and Vicente Angeles, G.R. No. 158693, November 17, 2004, Supreme Court of the Philippines, En Banc, LawPhil, https://lawphil.net/judjuris/juri2004/nov2004/gr_158693_2004.html. Supports: a dismissal for a valid just cause (abandonment) remains valid despite a due-process procedural defect, but the employer is liable for nominal damages for that defect. Status: verified official source.

[3] King of Kings Transport, Inc., Claire Dela Fuente, and Melissa Lim v. Santiago O. Mamac, G.R. No. 166208, June 29, 2007, Supreme Court of the Philippines, LawPhil, https://lawphil.net/judjuris/juri2007/jun2007/gr_166208_2007.html. Supports: the detailed content and timing requirements of the twin-notice rule (first notice specificity, minimum response period, hearing, second notice). Status: verified official source.

Administrative issuances / official agency materials

[6] National Privacy Commission, Advisory Opinion No. 2024-005, May 21, 2024, https://privacy.gov.ph/wp-content/uploads/2024/05/Advisory-Opinion-No.-2024-005.pdf. Supports: AI-based scoring of employee call/email performance data may rely on “legitimate interest” as a lawful DPA processing basis, subject to necessity, proportionality, and the employee’s right to object. Status: verified official source; advisory opinion addressed to a specific requesting party — persuasive/interpretive, not a generally binding rule for all employers.

[7] National Privacy Commission, Advisory No. 2024-04, “Guidelines on the Application of Republic Act No. 10173… to Artificial Intelligence Systems Processing Personal Data,” December 19, 2024, https://privacy.gov.ph/wp-content/uploads/2025/02/Advisory-2024.12.19-Guidelines-on-Artificial-Intelligence-w-SGD.pdf. Supports: transparency, accountability, fairness, data-minimization, mandatory human-intervention, and data-subject-rights requirements for AI systems processing personal data. Status: verified official source. Note: contains no employment-specific provisions; applied here by extension to the HR context.

Sources rechecked as of: August 20, 2026. Last materially reviewed: August 20, 2026.

Disclaimer

This article is provided for general educational and legal-information purposes only and does not constitute legal advice. Labor and data-privacy disputes depend on specific facts and on the law and agency guidance in effect at the relevant time, both of which can change. Nothing in this article guarantees a particular outcome for any hiring, termination, or privacy-related dispute. Readers facing an actual dispute should consult a Philippine labor lawyer, the Department of Labor and Employment (DOLE), the National Labor Relations Commission (NLRC), the National Conciliation and Mediation Board (NCMB), or the National Privacy Commission (NPC) as appropriate. LaborCode.ph is an independent legal-information platform and is not a law firm, government agency, tribunal, or official government website.

Similar Posts