Yes. An AI-generated threat can be reported under Philippine cybercrime laws when a person uses a computer system, social media account, messaging app, email, chatbot, bot, deepfake voice or video, or any other information and communications technology to threaten, intimidate, blackmail, harass, or coerce someone. Philippine law does not usually ask, “Did a human type every word?” The more important questions are: Who caused the threat to be sent or posted? What exactly was threatened? Was it communicated to the victim? Was ICT used? And is there evidence linking the online account, device, number, IP address, payment account, or identity to a real person?

As of July 2026, the Philippines does not need a separate “AI threats law” before an AI-assisted threat can be investigated. Existing laws — especially the Cybercrime Prevention Act of 2012, Republic Act No. 10175, the Revised Penal Code, and special laws on harassment, violence, privacy, and child protection — can already apply depending on the facts.

What Counts as an AI-Generated Threat?

An AI-generated threat is not limited to a message created by a chatbot. It can include:

A threatening message drafted with AI and sent through Messenger, Viber, Telegram, WhatsApp, email, SMS, X, TikTok, Instagram, or another platform.
A deepfake voice note saying, “I will hurt you” or “Pay me or I will harm your family.”
A fake AI-generated video showing a person threatening someone.
A bot or automated account repeatedly sending violent, sexual, or blackmail-style threats.
AI-generated images used to threaten, shame, extort, or sexually harass a victim.
A fake account using AI-generated photos, voice, or profile details to impersonate someone and intimidate another person.

Philippine criminal law still focuses on the human actor. AI is usually treated as a tool. The report is generally against the person who created, caused, controlled, sent, posted, distributed, or benefited from the threat — not against the AI software itself.

A threat is more likely to be actionable when it is specific, directed at a person, and capable of causing fear or intimidation. For example:

“I will kill you when you go home tonight.”
“Pay ₱50,000 or I will post your private photos.”
“I know where your child studies. I will hurt them if you report me.”
“I made a fake nude video of you. Send money or I will upload it.”

On the other hand, not every rude, angry, or offensive AI-generated statement is automatically a criminal threat. The exact words, context, identity of the sender, history between the parties, and surrounding acts matter.

Legal Basis: Why AI-Generated Threats Can Fall Under Cybercrime Law

Cybercrime Prevention Act: RA 10175

The main cybercrime law is Republic Act No. 10175, the Cybercrime Prevention Act of 2012.

The most important provision for AI-generated threats is Section 6. It states that crimes already punishable under the Revised Penal Code and special laws are also covered by RA 10175 when committed by, through, and with the use of information and communications technologies. The law also provides for a higher penalty in such cases.

This is why an old crime, such as threats under the Revised Penal Code, can become a cyber-related offense when committed through a phone, laptop, social media account, email, messaging app, fake profile, or AI-powered tool.

In simple terms: if the threat is a crime offline, using ICT to commit it can bring it within the Cybercrime Prevention Act.

Revised Penal Code: Grave Threats, Light Threats, and Other Light Threats

Threats are punished under the Revised Penal Code, especially:

Article 282 — Grave Threats
Article 283 — Light Threats
Article 285 — Other Light Threats
Article 286 — Grave Coercions
Article 287 — Light Coercions and unjust vexations

In Caluag v. People, G.R. No. 171511, March 4, 2009, the Supreme Court explained the difference among grave threats, light threats, and other light threats. A grave threat involves a threatened wrong that amounts to a crime, such as killing, serious physical injury, rape, arson, kidnapping, or other criminal harm. A light threat involves a threatened wrong that does not amount to a crime but is made with a condition. Other light threats cover lesser threatening acts not falling under the more serious categories.

The most common cyber-related threat situations are:

Situation	Possible legal issue
“I will kill you” sent by chat, email, or DM	Grave threats under Article 282, in relation to Section 6 of RA 10175
“Pay me or I will hurt your family”	Grave threats with a condition; possible extortion-related facts
“I will post your private photos unless you obey me”	Grave threats, coercion, Safe Spaces Act, Anti-Photo and Video Voyeurism Act, VAWC if relationship-based
Repeated AI-generated messages meant to frighten or harass	Threats, unjust vexation, cyber harassment-related offenses depending on facts
Fake AI voice pretending to be another person making threats	Threats, computer-related identity theft, forgery, or fraud depending on how it was made and used

The exact charge is determined by investigators and prosecutors after reviewing the complaint-affidavit and evidence.

Other Philippine Laws That May Apply

AI-generated threats often overlap with other laws. A single incident may involve more than one possible offense.

Cyber Libel

If the AI-generated content contains a public or shareable defamatory accusation — for example, falsely calling someone a criminal, sex offender, scammer, or adulterer — the case may also involve cyber libel under Section 4(c)(4) of RA 10175 in relation to Articles 353 and 355 of the Revised Penal Code.

The Supreme Court in Disini v. Secretary of Justice, G.R. No. 203335, February 11, 2014, upheld the constitutionality of cyber libel as to the original author of the libelous post, while striking down or limiting other portions of the Cybercrime Prevention Act. This case remains important because it explains how free speech, online expression, and cybercrime enforcement are balanced in the Philippines.

Safe Spaces Act: Online Gender-Based Sexual Harassment

If the AI-generated threat is sexual, misogynistic, homophobic, transphobic, or gender-based, the Safe Spaces Act, RA 11313, may apply.

Section 12 of RA 11313 expressly covers gender-based online sexual harassment, including acts using ICT to terrorize or intimidate victims through physical, psychological, and emotional threats; unwanted sexual remarks; cyberstalking; incessant messaging; impersonation; posting lies to harm reputation; and uploading or sharing sexual photos, voice, or videos without consent.

This is especially relevant for:

AI-generated nude images or sexual deepfakes.
Threats to upload intimate images.
Repeated sexualized threats through DMs.
Fake accounts impersonating a victim in a sexual manner.
Online stalking using AI-generated accounts or content.

VAWC: Threats by a Spouse, Ex, Partner, or Dating Partner

If the threat is made by a husband, former husband, live-in partner, ex-partner, boyfriend, former boyfriend, or someone with whom the woman has or had a sexual or dating relationship, the Anti-Violence Against Women and Their Children Act, RA 9262, may apply.

RA 9262 covers not only physical abuse but also psychological violence, including intimidation, harassment, stalking, public ridicule, repeated verbal abuse, and acts causing mental or emotional suffering. AI-generated threats sent by an abusive partner can support a VAWC complaint and may also support a request for a Barangay Protection Order, Temporary Protection Order, or Permanent Protection Order.

Child Protection and Online Sexual Abuse Laws

If the victim is a minor, the case becomes more serious. Depending on the facts, the following laws may apply:

AI-generated sexual images of minors, threats to distribute sexual material involving a child, grooming, coercion, or blackmail involving a minor should be treated as urgent.

Anti-Photo and Video Voyeurism Act

If the threat involves intimate images or videos, the Anti-Photo and Video Voyeurism Act, RA 9995, may apply.

This law is especially relevant when the offender threatens to publish, send, sell, or distribute private sexual images or videos. AI deepfakes may raise more complex proof issues, but if the content uses a person’s face, body, identity, or private image to shame or extort them, investigators may consider RA 9995 together with cybercrime, Safe Spaces Act, and other laws.

Civil Code Remedies

Even when a criminal case is difficult because the sender is anonymous or abroad, the victim may still have civil remedies if the wrongdoer is identified.

Relevant provisions include:

Civil Code Article 19 — every person must act with justice, give everyone their due, and observe honesty and good faith.
Article 20 — a person who causes damage contrary to law may be liable.
Article 21 — a person who wilfully causes loss or injury in a manner contrary to morals, good customs, or public policy may be liable.
Article 26 — protects dignity, personality, privacy, and peace of mind.
Article 32 — allows civil actions for violations of constitutional rights.
Article 33 — allows independent civil actions in certain cases such as defamation, fraud, and physical injuries.

Civil claims may include damages, injunctions, or other relief, but they require a separate court process and usually take longer than immediate police or platform action.

Step-by-Step: What to Do If You Receive an AI-Generated Threat

1. Treat Immediate Danger as an Emergency

If the threat mentions a specific time, place, weapon, child, address, workplace, school, or travel route, treat it seriously.

Practical steps:

Move to a safe place.
Inform a trusted person.
Contact local police, barangay officials, building security, school security, or workplace security.
Use emergency channels such as 911 when there is immediate danger.
If the threat involves domestic violence or VAWC, ask about a Barangay Protection Order at the barangay where the victim resides or where the incident occurred.

Do not wait for the cybercrime unit to identify the sender before addressing physical safety.

2. Preserve Evidence Before Blocking or Reporting the Account

Many victims immediately block the account or report it to the platform. That is understandable, but it can also make evidence harder to retrieve.

Before blocking, save:

Full screenshots showing the message, sender name, profile photo, date, and time.
The profile URL, username, handle, account ID, email address, phone number, or QR code.
The full conversation thread, not only the worst message.
Original image, audio, or video files, if available.
Message links, post links, comment links, group links, and page links.
Delivery receipts, call logs, email headers, and platform notifications.
Proof of demands for money, sexual acts, silence, apology, resignation, or any other condition.
Proof of payment requests, e-wallet numbers, bank accounts, crypto wallets, or remittance details.
Your own timeline of events.

For deepfake voice or video threats, keep the original file if possible. Do not repeatedly compress, edit, crop, or repost it. Investigators may need the file metadata, upload source, and platform records.

3. Do Not Hack, Entrap Alone, or Secretly Record Private Calls

Preserving evidence is different from committing another offense.

Avoid:

Guessing or stealing passwords.
Logging into the suspect’s account without authority.
Installing spyware.
Pretending to be law enforcement.
Threatening the sender back.
Secretly recording private calls without proper legal guidance.

The Anti-Wiretapping Law, RA 4200, can create problems when private communications are secretly recorded without the required consent or legal authority. If the threat is made through a call, preserve call logs, voicemails, messages, and written details, then coordinate with law enforcement on the proper way to document future communications.

4. File a Report With the Proper Office

You may report cyber-related threats to:

Office	When it is commonly used
PNP Anti-Cybercrime Group (PNP ACG)	Online threats, fake accounts, cyber harassment, cyber libel, identity theft, digital evidence investigation
NBI Cybercrime Division (NBI CCD)	Cybercrime complaints requiring investigation, digital forensics, subpoenas, coordination with platforms
Local police station	Immediate danger, blotter, physical safety, threats by known persons, urgent response
Barangay	Blotter, local safety intervention, VAWC Barangay Protection Order, neighborhood incidents
City or Provincial Prosecutor’s Office	Filing of criminal complaint-affidavit for preliminary investigation
School, employer, platform, or building security	Parallel administrative or safety action, especially for students, employees, tenants, or workplace harassment
DOJ Office of Cybercrime	Cybercrime policy, coordination, and international cybercrime matters, especially when cross-border assistance is needed

The NBI Citizens’ Charter page for computer crime complaints states that complainants may proceed to the Cybercrime Division to file a complaint or request for investigation, undergo preliminary interview and initial investigation, execute sworn statements or submit prepared affidavits, and present devices relevant to the probe. It also indicates no fee for the listed initial steps and an estimated total processing time of about 1 hour and 10 minutes for the initial transaction.

In practice, the initial intake may be same-day, but identifying an anonymous sender can take much longer.

5. Prepare a Clear Complaint-Affidavit

A complaint-affidavit is a sworn written statement narrating what happened. It should be factual and organized.

Include:

Your full name, address, contact details, and valid ID.
The suspect’s name, username, account, phone number, email, or any known identifying details.
The exact words or content of the threat.
The date, time, platform, and device used.
How you received or discovered the threat.
Why you believe the account belongs to the suspect, if known.
Any prior incidents, relationship history, workplace issue, family dispute, debt issue, romantic history, or extortion demand.
The effect on you: fear, inability to go to work or school, anxiety, changed routine, security costs, reputational harm, or financial loss.
A list of attached screenshots, files, links, and witness statements.
A request that the matter be investigated for the appropriate offense.

If you are abroad, your affidavit may need notarization before a Philippine consulate or authentication/apostille depending on where it is executed and how it will be used in the Philippines.

6. Expect Investigation, Subpoenas, and Possible Digital Forensics

For anonymous or fake accounts, investigators may look for:

IP logs.
Account registration details.
Recovery email or phone number.
Device identifiers.
Payment or e-wallet records.
SIM registration details.
Platform preservation data.
CCTV or physical-world corroboration.
Links between the online account and the suspect’s known accounts.

This is where many cases slow down. Platforms may be based abroad. Some data may be deleted quickly. Some accounts use VPNs, fake SIMs, or stolen photos. This does not make the case impossible, but it means early evidence preservation is important.

The DOJ Office of Cybercrime is the central authority for international mutual assistance and extradition in cybercrime and cyber-related matters. For foreign platforms or suspects abroad, investigators may need formal legal processes, not just screenshots.

7. Follow Through With the Prosecutor’s Office

After investigation, the complaint may be referred to the City or Provincial Prosecutor’s Office for preliminary investigation. The prosecutor evaluates whether there is probable cause to file the case in court.

Possible outcomes include:

The complaint is dismissed for lack of evidence.
The prosecutor requires more evidence or clarification.
The prosecutor files an Information in court.
The case proceeds to arraignment, pre-trial, trial, and judgment.
The parties may have parallel civil, administrative, workplace, school, or platform remedies.

Cybercrime cases are not always fast. A simple complaint involving a known sender may move faster. A fake-account case requiring platform data, foreign cooperation, or forensic examination may take months or longer.

Evidence Checklist for AI-Generated Threats

Evidence	Why it matters
Screenshots with date and time	Shows the exact threat and context
Profile URL, username, account ID, phone number, email	Helps identify the sender
Full chat thread	Prevents claims that the message was taken out of context
Original audio, video, or image file	Important for deepfake or AI-generated media analysis
Screen recording showing the profile and messages	Helps prove the content existed on the platform
Device used to receive the threat	May be examined by investigators
Witness statements	Supports fear, identity, or surrounding facts
Payment demands or account numbers	Useful in blackmail or extortion-related facts
Platform report receipts	Shows steps taken and timing
Medical, psychological, school, or work records	Supports harm or impact when relevant
Barangay or police blotter	Helps establish timely reporting

The Supreme Court has recognized that chat logs, videos, photos, and messages may be used as evidence in criminal cases when properly obtained and authenticated. In People v. Rodriguez, G.R. No. 263603, October 9, 2024, the Supreme Court sustained the use of online chat logs and videos in a criminal case. In another Supreme Court ruling involving Facebook Messenger evidence, the Court also held that photos and messages obtained by private individuals may be admissible when the facts do not show unlawful state intrusion.

The practical lesson is simple: screenshots help, but they are stronger when supported by original files, links, device access, testimony, and a clear chain of how the evidence was obtained.

Common Problems in AI-Generated Threat Cases

“The sender said it was just AI, not serious.”

That does not automatically excuse the act. If a person intentionally sends an AI-generated death threat, blackmail message, or sexual threat to scare someone, the use of AI may be treated as the method, not a defense.

The sender’s intent and the victim’s reasonable fear are assessed from the whole context.

“The account is fake.”

Fake accounts are common in cybercrime complaints. A fake name does not prevent reporting, but it does make evidence gathering more important. Preserve URLs, profile IDs, timestamps, message IDs, and any link to known phone numbers, emails, photos, writing style, payment accounts, or prior conversations.

“The threat was deleted.”

Deleted content may still be supported by screenshots, notifications, witness testimony, device records, backups, and platform data. Report quickly because platforms may retain certain records only for limited periods.

“The offender is abroad.”

Cross-border cybercrime cases are harder but not hopeless. If the victim is in the Philippines, the damage occurred in the Philippines, the suspect is Filipino, the computer system used has links to the Philippines, or Philippine accounts and services were involved, local authorities may still evaluate the complaint. International cooperation may be needed through proper government channels.

“The victim is a foreigner.”

Foreigners in the Philippines may report threats to Philippine authorities. A foreigner abroad dealing with a Philippine-based offender may also coordinate with Philippine law enforcement, but affidavits and foreign documents may need consular notarization, apostille, or proper authentication before use in Philippine proceedings.

“The threat involves a workplace or school.”

A criminal complaint may proceed separately from workplace or school remedies. Under RA 11313, employers and schools have duties to prevent and address gender-based sexual harassment, including technology-based conduct. A workplace or school complaint can help with safety measures, but it does not replace a cybercrime or criminal complaint when the conduct is serious.

Frequently Asked Questions

Can I report an AI-generated death threat to the PNP or NBI?

Yes. If the threat was sent or posted through ICT, you may report it to the PNP Anti-Cybercrime Group, NBI Cybercrime Division, local police, or prosecutor’s office. A death threat may fall under grave threats under Article 282 of the Revised Penal Code, in relation to Section 6 of RA 10175.

Is it still a crime if the message was written by ChatGPT or another AI tool?

It can still be a crime if a person used the AI-generated message to threaten, intimidate, harass, coerce, blackmail, or defame someone. Philippine law generally looks at the human who caused the message to be created, sent, posted, or used.

Can a deepfake voice threat be reported?

Yes. A deepfake voice note or video threatening harm may be reported. Preserve the original file, platform link, sender profile, timestamps, and surrounding messages. The case may involve threats, identity-related offenses, cyber harassment, or other laws depending on the facts.

Are screenshots enough to file a cybercrime complaint?

Screenshots are useful, but they are stronger with supporting evidence such as URLs, account IDs, original files, device access, screen recordings, witnesses, and a sworn affidavit. Do not rely only on cropped screenshots if you can preserve the full context.

What if the AI-generated threat was sent by an anonymous or dummy account?

You can still report it. Investigators may seek account records, IP data, device information, SIM registration links, payment records, or platform data. Anonymous cases are more difficult, so early preservation of evidence is important.

Can I post the threat publicly to expose the sender?

Be careful. Publicly posting the threat may spread harmful content, expose private information, affect minors, or create defamation and privacy issues. Preserve the evidence first and report through proper channels. If public safety requires warning others, avoid unnecessary personal details and coordinate with authorities when possible.

Can I block the sender after saving evidence?

Yes. After saving the evidence, blocking may be appropriate for safety and mental well-being. If there is immediate danger, also inform police, barangay officials, workplace or school security, or trusted people who can help protect you.

What if the threat is sexual or involves AI-generated nude images?

Report it as soon as possible. The case may involve RA 11313, RA 9995, RA 10175, RA 9262 if relationship-based, and child protection laws if the victim is a minor. Preserve the content without reposting or forwarding it unnecessarily.

Can foreigners report AI-generated threats in the Philippines?

Yes. Foreigners in the Philippines may report to local police, PNP ACG, NBI CCD, or the prosecutor. Foreigners abroad may need authenticated affidavits and may need to coordinate with Philippine authorities, especially if the suspect or harmful effect is connected to the Philippines.

How long does a cybercrime threat case take?

The initial complaint may be received the same day. The NBI Citizens’ Charter lists the initial computer crime complaint process as having no fee and an estimated initial transaction time of about 1 hour and 10 minutes. But investigation, identification of anonymous accounts, subpoenas, platform cooperation, preliminary investigation, and court proceedings can take months or longer.

Key Takeaways

AI-generated threats can be reported under Philippine cybercrime laws when ICT is used to create, send, post, or distribute the threat.
The main legal basis is Section 6 of RA 10175, which covers Revised Penal Code crimes and special-law offenses committed through ICT.
Threats may fall under grave threats, light threats, coercion, cyber libel, online sexual harassment, VAWC, child protection laws, privacy laws, or civil damages, depending on the facts.
Philippine law focuses on the person who used or controlled the AI tool, not the AI system itself.
Preserve evidence before blocking or reporting the account: screenshots, URLs, usernames, original files, timestamps, devices, and witness details.
For immediate danger, prioritize safety through police, barangay, school, workplace, or security assistance.
Anonymous or foreign-based accounts are harder to investigate, but early reporting helps preserve platform and technical evidence.
Screenshots help, but the strongest complaints include a clear sworn affidavit, original digital files, full context, and proof connecting the online act to a real person.