Artificial intelligence is changing how Philippine lawyers research, review documents, draft working copies and manage cases. It can shorten a first pass through a large file, but a fluent answer can still contain a false case citation, miss a controlling exception or expose client information. The practical rule is simple: use AI to assist a defined legal task, then have a competent lawyer verify the sources, facts, reasoning and final advice. For a step-by-step check before filing, see our guide to verifying AI-assisted pleadings and catching hallucinated citations.
The governing context includes the Supreme Court’s Code of Professional Responsibility and Accountability (CPRA), the Data Privacy Act and NPC Advisory No. 2024-04 when personal data are processed, and any applicable court rules and client obligations. The Supreme Court’s 2026 AI governance framework governs AI use in the Judiciary and expressly includes court users within its stated scope. Check its requirements when using AI in that setting; it is not a blanket rule for every private law-office task.
Evidence and action
- Rule/standard: The CPRA governs lawyers’ professional conduct; the DPA and NPC guidance apply to relevant processing of personal data. The Supreme Court has also adopted a human-centered AI framework for its judicial system, including court users in its stated scope.
- Qualification: An AI output is neither a verified Philippine legal authority nor a substitute for a lawyer’s independent professional judgment. Whether a particular AI use is suitable depends on the data, client instructions, court requirements and product terms.
- Risk/evidence: Keep the authoritative case or statute, its date and status, the source document used, and a record of the human review. Restrict confidential client files and personal data until the tool and agreement have been assessed.
- Timing: Check controlling law and procedural deadlines as they apply to the actual case. There is no universal AI-specific filing deadline for all legal work.
- Next step: Start with a narrow, low-risk workflow using public or properly de-identified material. Test errors and review obligations before expanding access.
Where is AI already useful in legal practice?
AI is most useful when the task has a clear input, a checkable output and a person accountable for accepting it. A tool might summarize a long contract or surface search terms from discovery material. It cannot establish, merely by sounding certain, that a cited decision exists or that a proposed argument fits the client’s facts.
| Task | Useful first pass | What a lawyer must verify |
|---|---|---|
| Legal research | Generate issues, search terms and a list of possible authorities | Obtain each statute, regulation or decision from an authoritative source; check the holding, date, jurisdiction and later treatment |
| Drafting | Prepare an outline, clause alternatives or a first draft | Match the client’s instructions, actual facts, procedural rules, evidence and intended legal effect |
| Contract review | Flag differences between versions and extract obligations | Read the operative text, schedules, defined terms, governing law and commercial context |
| Discovery and evidence | Organize documents and suggest chronology | Preserve originals, provenance and lawful access; verify the timeline against the record |
| Client intake | Sort questions and identify missing documents | Confirm facts directly, assess conflicts, avoid collecting more personal data than needed |
| Court administration | Help organize public rules or permissible filing workflows | Check the court’s current requirements and the status of any electronic submission |
These are potential uses, not evidence that every tool performs each task securely or accurately. Set acceptance criteria before piloting a product.
Does AI replace a lawyer in the Philippines?
AI can perform parts of the work, but it cannot carry a lawyer’s professional responsibility or independently represent a client merely because it can write a plausible pleading. Under the CPRA, the responsible lawyer’s duties of fidelity, competence, diligence and accountability remain relevant to an AI-assisted service. The Supreme Court has urged lawyers to verify factual claims and citations, safeguard confidential information and preserve independent reasoning. CPRA · Supreme Court statement on AI and human judgment
AI will likely change the mix of time spent on research, drafting and review. Whether that lowers a client’s bill, changes staffing or improves access to justice depends on the firm’s service model, error rate, oversight cost and client needs. It is a forecast, not an established legal outcome. For career paths, CyberCode’s Juris Doctor guide and forthcoming law-specialties guide address different questions.
Can a lawyer put client documents into an AI tool?
It depends on the document, the tool and the engagement. A public decision is different from an unfiled affidavit, a medical record or a privileged strategy memorandum. Before uploading client material, determine who is authorized to process it; what purpose the tool serves; whether the material contains personal or sensitive personal information; what the provider retains, accesses or shares; what the client agreement requires; and whether safer de-identification or a restricted deployment is possible.
NPC Advisory No. 2024-04 explains that the DPA, its implementing rules and NPC issuances apply when AI development or deployment processes personal data, including training and testing. It addresses lawful basis, transparency, accountability, data minimization, security and data-subject rights. That guidance concerns personal data; professional confidentiality and any applicable privilege or client undertaking present additional questions. A vendor claim that data are “private” is insufficient without checking the applicable plan, settings and contract. NPC Advisory No. 2024-04, sections 1–3 · Supreme Court on lawyers’ AI use
Example: A firm wants an AI assistant to summarize 500 pages of employment records. First separate public authorities from client-provided material; identify employee information; check client instructions, the vendor agreement and access controls; pilot with synthetic or adequately de-identified samples; then assign a lawyer to verify the citations, chronology and omissions against the originals. Do not assume that replacing names alone removes every identifier.
How should lawyers check AI-generated case citations?
Open the actual decision in the Supreme Court’s official publication or another reliable legal source. Confirm that the docket number, parties, date and tribunal match; read the relevant passage and distinguish the holding from a general observation; check whether a later case, statute or rule changed its effect; and confirm that the facts support the proposed use. Repeat the check for statutes, amendments and procedural rules.
Keep a short source record for each proposition that matters to the advice or pleading: proposition, authority, pinpoint, status checked, date accessed and reviewer. If an authority cannot be found or does not say what the draft claims, remove it. This is an editorial verification workflow based on the lawyer’s duty to deliver competent work; it is not a claim that one particular spreadsheet format is mandated by the CPRA. CPRA · Supreme Court guidance on verifying AI citations
What has the Philippine Supreme Court actually done on AI?
In February 2026, the Supreme Court adopted a Governance Framework on the Use of Human-Centered Augmented Intelligence in the Judiciary under A.M. No. 25-11-28-SC. Its stated principles include fairness, accountability and transparency in the judicial system. According to the Court’s announcement, its scope includes justices, judges, court officials and employees, court users, and relevant vendors or contractors. The announcement also describes approval and disclosure controls for AI used in court processes. A lawyer who is a court user should check the applicable framework and court instructions before using AI for that work. This differs from treating every private-office draft as a court-issued document. The CPRA separately governs lawyers, while NPC guidance addresses AI systems processing personal data.
The Court’s public statements emphasize human judgment and caution against unmindful use. Those statements help explain its policy direction; a speech is not itself a new statute or a universal permission to submit AI-produced material without checking it. Supreme Court framework announcement · September 2026 remarks
A practical AI policy for a Philippine law office
Write a one-page policy that staff can follow during a busy case. Begin with the following decisions:
- Allow specific tasks. Define what may be drafted, summarized or classified, and which work needs express partner or client authorization.
- Classify the input. Separate public material, internal work product, confidential client information and sensitive personal data.
- Approve tools and contracts. Record the provider, permitted users, retention and training terms, location or transfer issues where relevant, security controls and deletion route.
- Limit access. Give the tool only the files and actions needed for that task; require approval for sending, filing, deleting or changing a live record.
- Require source review. A named lawyer checks every material proposition and citation against the primary source and the case file.
- Keep a useful record. Preserve the task brief, approved sources, material corrections and final human sign-off, consistent with confidentiality and retention obligations.
- Test incidents. Know how to revoke access, preserve evidence and assess whether a data or security incident triggers internal, client, contractual or NPC action.
The list is a proposed office workflow, not a verbatim list of legal duties imposed on every firm. Tailor it to the practice area, client contracts, court rules and the data actually involved.
What will change for junior lawyers and clients?
Junior lawyers may spend less time producing a first draft and more time assessing whether it is right. Training should therefore include primary-source research, distinguishing holdings, handling evidence, interviewing clients, explaining uncertainty and checking an automated workflow. A firm that removes every research task from junior lawyers risks weakening the very expertise needed to catch AI errors. This is a practical training inference, not a measured prediction about employment.
Clients can reasonably ask what systems touch their information, who reviews the output and how errors are corrected. They should expect a clear scope of work and accountable advice. They should not assume that a cheaper-looking automated draft is complete legal advice for their circumstances.
FAQ
May a Philippine lawyer use ChatGPT or another AI model to draft a pleading?
AI may help prepare a working draft, but the responsible lawyer must check the facts, authorities, applicable rules and final filing. Confidentiality, data protection, client instructions and any court-specific requirement can change which tool or material may be used. CPRA · NPC AI advisory
Is a client’s consent always required before a firm uses AI?
There is no sound one-word answer for every use. Client confidentiality and engagement terms must be assessed; when personal data are processed, the firm must identify an appropriate lawful basis and meet applicable DPA duties. Consent is not the only possible lawful basis under the DPA, and a client’s general consent does not cure insecure or disproportionate processing. NPC Advisory No. 2024-04
Does the 2026 Supreme Court AI framework apply to private law firms?
It addresses AI use in the Judiciary and includes court users within its stated scope. A private lawyer using AI in court-related processes should check its applicable controls; other law-office uses still call for the CPRA, applicable privacy rules, client instructions and contractual commitments. Supreme Court framework announcement
Which legal tasks should remain human-led?
Client advice, strategic decisions, factual evaluation, witness credibility, signing and filing, and decisions that affect a person’s rights need accountable human judgment. An AI tool can support parts of the process, but the lawyer should be able to explain and defend the final work. Supreme Court on human judgment
CyberCode takeaway
The legal profession gains the most from AI when it makes source review, document organization and client service more effective. The measure of success is a defensible result: the right authority, the right facts, protected client information and an identifiable professional who checked the work.
Related CyberCode reading: Astra for Law in the Philippines explores one product and lawyers’ use cases. NPC Advisory No. 2024-04: AI Systems and Data Privacy explains the privacy framework. Juris Doctor in the Philippines covers the education and career route.
Featured image: Photo by Vitaly Gariev via Unsplash.