Philippines - AI Regulation for Elections (HB 10567)

Overview

House Bill No. 10567, the 'Deepfake Accountability and Transparency Act,' establishes regulatory framework for AI-generated manipulated media ahead of the May 2025 Philippine elections and beyond. Introduced July 1, 2024, the bill responds to cybersecurity and electoral integrity threats posed by deepfakes—digitally altered images, audio, and video that falsely depict individuals' speech or conduct using artificial intelligence. The legislation requires mandatory disclosure for all deepfake content, establishes Department of Information and Communications Technology (DICT) oversight coordinated with COMELEC, and imposes significant penalties (up to PHP 5 million) for non-compliance. The bill addresses concerns that generative AI technologies enable creation of convincing fake media capable of misleading voters, damaging reputations, and undermining democratic discourse, necessitating transparency mechanisms enabling individuals to distinguish authentic from manipulated content. House Bill 10567 complements COMELEC Resolution No. 11064's restrictions on AI-generated electoral content, creating layered protections. The bill remains pending in the House Committee on Information and Communications Technology, with enactment urgency driven by 2025 election timeline requiring implementation before campaign periods begin.

Definitions

Deepfake means any form of media including images, audio recordings, or video that falsely depicts a person's speech, conduct, appearance, or participation in events through use of artificial intelligence, machine learning algorithms, or generative AI systems, creating realistic but inauthentic representations. Generative AI Systems encompasses technologies capable of creating synthetic media including text-to-speech synthesis, voice cloning, face-swapping, video synthesis, image generation, and other AI techniques producing realistic fake content. Disclosure refers to clear, conspicuous, and persistent statements indicating that media contains AI-generated or digitally altered material, presented in formats enabling viewers/listeners to recognize manipulation. Creator means individuals or entities that generate deepfake content using AI technologies. Distributor includes individuals, organizations, or platforms that share, publish, broadcast, or otherwise make deepfake content available to audiences. Electoral Deepfake specifically refers to deepfakes depicting candidates, political parties, electoral processes, or voting-related information during campaign periods. Legitimate Use encompasses satirical, artistic, educational, or journalistic applications of AI media manipulation where context or explicit labeling makes clear the content is not factual representation.

Governance and Institutional Framework

The Department of Information and Communications Technology (DICT) serves as primary regulatory authority with responsibilities including developing implementing regulations specifying technical disclosure standards, visual label formats, and platform compliance procedures; monitoring online platforms, social media, broadcast media, and other communication channels for non-compliant deepfakes; coordinating with Commission on Elections (COMELEC) on electoral deepfakes, aligning enforcement with COMELEC Resolution No. 11064; investigating reported violations and gathering evidence for enforcement actions; imposing administrative penalties within regulatory authority; and referring serious violations for criminal prosecution. COMELEC coordinates on electoral applications through joint monitoring during campaign periods, expedited review of political deepfakes, and integrated enforcement combining Bill 10567 disclosure requirements with Resolution 11064 content prohibitions. The National Privacy Commission (NPC) may coordinate where deepfakes involve unauthorized use of personal data or biometric information. The Department of Justice handles criminal prosecutions for deepfake-related offenses under existing laws including cyber libel, fraud, or election crimes. Platform providers (Facebook, TikTok, YouTube, etc.) bear responsibilities for implementing disclosure requirement compliance mechanisms, content flagging systems, and cooperation with DICT investigations.

Key Focus Areas

First, mandatory disclosure requirements: All deepfake content must include clear verbal announcements (for audio/video) stating 'This content contains AI-generated or digitally altered material' or equivalent language, and persistent written/visual labels displayed throughout media consumption in fonts and contrasts ensuring readability. Disclosures cannot be obscured, minimized, or hidden through design choices. Second, electoral protection: Enhanced requirements and expedited enforcement for political deepfakes during campaign periods, protecting election integrity and voter decision-making. Third, platform accountability: Online platforms and media outlets distributing deepfakes share responsibility for ensuring compliance, implementing detection systems and disclosure verification. Fourth, rapid response mechanisms: Expedited takedown procedures for non-compliant deepfakes, particularly electoral deepfakes, enabling swift action against manipulation attempts. Fifth, public education: DICT campaigns informing public about deepfakes, disclosure requirements, and media literacy skills for identifying manipulation.

Implementation Framework

Implementation follows urgent timeline given May 2025 elections: Phase 1 (Enactment to Month 3): DICT develops implementing regulations through consultation with technology companies, media organizations, electoral experts, and civil society; publishes technical standards for disclosure formats; establishes monitoring infrastructure and reporting mechanisms. Phase 2 (Months 3-6): Public education campaigns launched nationwide explaining disclosure requirements and recognition of deepfakes; platform providers implement compliance systems; voluntary compliance encouraged before mandatory enforcement. Phase 3 (Month 6+): Mandatory compliance begins with active monitoring and enforcement; heightened scrutiny during campaign periods (February-May 2025); ongoing refinement based on implementation experience.

Monitoring and Evaluation

DICT operates monitoring systems including automated scanning of major platforms for deepfake content; public reporting mechanisms enabling individuals to flag non-compliant deepfakes; coordination with platform providers' internal detection systems; and periodic audits of compliance rates and enforcement effectiveness. Metrics tracked include number of deepfakes detected, disclosure compliance rates, enforcement actions taken, public awareness levels, and impact on electoral integrity indicators.

Penalties, Liability, and Appeals

Penalties include fines up to PHP 5 million (approximately USD 90,000) for creators and distributors failing to disclose deepfake nature of content. Enhanced penalties for subsequent violations or aggravated cases involving electoral deepfakes, targeting vulnerable individuals, or causing significant harm. Content takedown orders requiring immediate removal of non-compliant deepfakes. Potential criminal prosecution under Cybercrime Prevention Act (cyber libel, computer-related fraud), Revised Penal Code (defamation), or Omnibus Election Code (election offenses) depending on deepfake use. Civil liability enabling individuals depicted in harmful deepfakes to seek damages. Appeals allow parties subject to DICT penalties to contest determinations through administrative review and ultimately judicial review in Philippine courts.

Relationship to Other Instruments

House Bill 10567 directly complements COMELEC Resolution No. 11064 (September 2024) which prohibits AI-generated deepfakes in political campaigns for 2025 elections. While Resolution 11064 focuses on prohibition and electoral context, Bill 10567 establishes broader disclosure-based regulation applicable beyond elections. The bill coordinates with House Bill No. 7396 (AI Development Act) and House Bill No. 7913 (AI Bill of Rights), which address general AI governance and may be consolidated. The bill aligns with the Data Privacy Act of 2012 regarding unauthorized use of personal data and biometric information in deepfakes. The Cybercrime Prevention Act of 2012 provides criminal liability framework for deepfake-enabled offenses. Internationally, the bill reflects approaches in the EU's proposed AI Act provisions on transparency for synthetic media, U.S. state-level deepfake disclosure laws, and emerging global consensus on combating AI-enabled disinformation.

International Alignment

House Bill 10567 aligns with emerging international standards for deepfake regulation. The disclosure-based approach mirrors provisions in the EU AI Act requiring transparency for AI-generated content capable of misleading individuals. The bill reflects principles from UNESCO's guidelines on information integrity emphasizing transparency and media literacy in combating disinformation. Electoral protection provisions align with international election observation standards and democratic integrity frameworks. The Philippines joins jurisdictions including several U.S. states, EU members, and others implementing deepfake disclosure requirements, contributing to global norm development while adapting regulations to Philippine electoral and cultural contexts.

Implementation Timeline

Sources and References