1 hour ago
1
Upgrade to High-Speed Internet for only ₱1499/month!
Enjoy up to 100 Mbps fiber broadband, perfect for browsing, streaming, and gaming.
Visit Suniway.ph to learn
MANILA, Philippines – Deepfakes or synthetic media generated by artificial intelligence (AI) and machine learning to mimic a person’s appearance, voice, or actions have evolved from a technical curiosity into a potent tool for disinformation, fraud, and harassment.
In the Philippines, high-profile victims like President Bongbong Marcos Jr. and public figures including Rappler CEO and Nobel Peace Prize laureate Maria Ressa, Senator Bam Aquino, businessmen Ramon Ang, former senator Manny Villar, and other prominent journalists such as Karen Davila and Karmina Constantino have already been targeted by manipulated content that is misleading to the public and may potentially cause reputational damage.
Deepfakes can also be used to make sexual images that target women, as recently seen in incidents involving the AI tool Grok. Celebrities such as Angel Aquino have also been targeted as subjects in deepfake porn. “It was dehumanizing,” Aquino said, in a Senate hearing in September 2025.
Senator Risa Hontiveros, at the same hearing, warned, “The use of technology in perpetrating harmful behaviors, abuses, and crimes against women and children are not new; but the potential scale in the creation of AI-generated non-consensual and abusive materials is unprecedented.”
In response to those, the Philippines’ legislative bodies have crafted bills looking to make the use of deepfake technology in a malicious manner unlawful.
In the 20th Congress, these bills are House Bills (HB) 250, 2312, 3214, 3425, 3823, 5558, HB 6539, 6632, 7584, 7903, and 8773, and Senate Bills 758 and 1714.
Amendments to the Safe Spaces Act or Bawal Bastos Act were passed in January 2025, addressing AI-based sexual abuse, but the new bills attempt to tackle deepfakes more comprehensively.
HB 2312 says, “While AI holds vast potential to drive innovation and national progress, the malicious use of deepfake technology poses significant and evolving risks to individual privacy, public trust, and democratic integrity.”
Consent, disclosure, exemptions for parody, artistic expression
Several core pillars are consistent across the proposed measures, discussed below:
- Consent
Every bill mandates that the creation, distribution, or commercial use of a deepfake depicting an identifiable individual requires a form of consent that is written and informed. This principle ensures that individuals maintain exclusive control over their own digital likeness.
HB 6632 or the “Personal Identity Right Act” are among the bills highlighting this. It is anchored on the principle that a person’s identity is their property and that they “shall have the exclusive control and personal identity right over their name, face, voice, and likeness, which is effective upon birth and extending until fifty (50) years after their death.”
It specifies that any deepfake use for “any purpose without a valid written consent” which “must indicate the specific and legitimate intended use,” and is revocable by the owner any time.
The two Senate bills emphasize as well the rights to one’s likeness and identity. These are to be protected “against the use of such attributes by artificial intelligence, machine learning models, or other technologies to create, publish, or distribute synthetic or altered content, including deepfakes, without the person’s express consent,” the bills say.
- Exemptions for non-malicious uses
The bills do provide exemptions for uses such as satire, parody, news reporting, artistic expression, and educational purposes, provided that the creator or the publisher has sufficient disclosure informing the audience that the material is AI-generated or AI-manipulated.
The artificial content must also not be harmful or does not constitute malicious use towards the subject.
HB 6632 states “the use of deepfake media is deemed non-infringing [on a person’s identity rights] only if it does not constitute malicious use.”
HB 5558 however looks at harms even under the said exemptions. It says: Exemptions for cases involving “parody, satire, caricature, or social critique” will be “voided if the imitation constitutes misinformation that may cause concrete harm to others’ rights or significant interests, such as health, privacy, reputation, or property.”
- Mandatory disclosure
To prevent public confusion, the bills require all synthetic media to carry a clear and conspicuous disclosure. For visual media, this typically takes the form of a prominent watermark or text overlay, while audio recordings must include a verbal disclaimer at the beginning.
HB 7903 or the “Deepfake Accountability and Transparency Act” proposes some of the heaviest fines with regards to the failure of disclosures, highlighting the importance of making people aware that what they’re consuming is deepfake content.
Those who “create or distribute a harmful deepfake without proper disclosure,” can face fines ranging from P2 million to P5 million, and imprisonment of 6 to 12 years.
It proposes bigger fines of up to P10 million and 20 years in jail for those who “use deepfake technology to defame, deceive, commit fraud, or incite violence” or “obtain consent through coercion”
Both Senate bills 758 and 1714, provide clearer details for disclosures. For audiovisual content, the bills mandate at least one clearly articulated verbal statement identifying the content as AI-generated, accompanied by a concise description of the extent of the alteration. Simultaneously, these clips must feature an unobscured written statement in readable text appearing at the top of the image throughout the entire duration of the visual element. Furthermore, a visible link, icon, or similar tool must be present to signal that the content is a product of generative artificial intelligence.
The requirements for other media types follow a similarly strict logic. Visual-only content must include the same continuous written text at the top of the image and a clearly visible icon or link.
For audio-only content, a verbal statement identifying the AI-generated elements and describing the alterations is required at the beginning of the record. Notably, if an audio recording exceeds two minutes in length, the bills require that an additional verbal statement and description be repeated at some interval during each two-minute period thereafter.
- Defining malicious use
The measures define “malicious deepfakes” as those created with an intent to deceive for personal gain, defraud, incite violence, and defame to name some.
HB 3425 or “The Anti-Malicious Deepfake Act of 2025” describes other types of malicious deepfakes:
- Non-consensual intimate deepfakes, which “depicts an identifiable individual engaging in sexually explicit conduct, nudity, or other intimate acts without their consent, regardless of whether the depicted individual is a minor or an adult”
- Political disinformation deepfakes, which depicts candidates in an election or a public official, with “the intent to injure the candidate’s or public official’s reputation or to deceive voters or the public, within sixty (60) days of an election.”
- Harassment, stalking, or blackmail deepfakes
- Interagency coordination
The Department of Information and Communications Technology (DICT) and the Department of Justice (DOJ) are consistently designated as the lead agencies for enforcement and rulemaking, often in coordination with the National Privacy Commission (NPC).
HB 8773 or the “Deepfake Digital Content Accountability and Enforcement Act” is unique in that it proposes that “authorized deputies of the Movie and Television Review and Classification Board shall assist in monitoring publicly accessible digital platforms for suspected deepfake content.”
Novel aspects
One of the more innovative approaches is found in HB 03214 or the “Deepfake Regulation Act,” which allows Filipinos to register their face, body, and voice as a trademark under a dedicated office at the Intellectual Property Office of the Philippines (IPOPHL). But it does make clear that Filipinos have a right to their own likeness without prior registration. The bill doesn’t state what the benefits might be to registration.
Similarly, HB 5558 (“Deepfake Regulation Act”) and HB 6632 seek to protect deepfakes and AI-generated simulations as copyrightable subject matter under the Intellectual Property Code or RA 8293.
As mentioned above, HB 6632 gives posthumous rights to people, extending their right to their likeness 50 years after death. This means that even after death, a person has the right to their likeness and the right to protect their likeness from deepfakes, as managed by their estate or surviving kin. HB 6632 is also the only bill to explicitly state that an individual has the right to revoke their consent for the use of their likeness.
While several bills require platforms to have accessible reporting mechanisms for deepfake abuse, as well as content takedown windows ranging from 24 hours to 28 hours, SB 758, SB 1714, and HB 5558 also require platforms to maintain anonymized public logs of all takedown actions and reporting data for transparency.
HB 6539 or the “Anti-Deepfake Abuse and Digital Integrity Act” has a provision requiring internet service providers to install available technology, program, or software that will block internet access or filter all websites relating to the creation and use of unlawful deepfake materials as provided in this Act.” Historically, the DICT has threatened to block apps such as Grok for its use in deepfake naked images.
HB 3425 addresses the borderless nature of internet content by defining “extraterritorial jurisdiction.”
“The provisions of this Act shall apply to any person who commits any of the prohibited acts outside the territorial jurisdiction of the Philippines, provided that the malicious deepfake targets or causes harm to a Filipino citizen or resident, or is intended to affect public discourse or elections within the Philippines. The Department of Justice, in coordination with relevant international bodies, shall take necessary steps to pursue such offenders,” it states.
Corporations may also face penalties for violations under HB 7584 and HB 6632 if a deepfake offense is committed as an employee working in a company or because the company failed to supervise an employee. – Rappler.com
