Benchmarking AI Regulations: Lessons from the EU AI Act for the Philippines
A comparative analysis between the EU AI Act and Philippine AI bills
Hello there!
Today, I will go deep, very deep into the current state of upcoming AI regulatory bills in the Philippine Congress as compared to the EU AI Act. This is part of a research paper I did for my AI Ethics class in graduate school. So brew a cup of coffee, take a seat, get comfy, and enjoy the ride.
Introduction
On August 1, 2024, the EU AI Act was enacted after years of deliberation. It is the first ever regulation on AI in the world, and may soon become the benchmark for global AI regulation efforts, following the footsteps of GDPR, which is among the most influential data protection laws in the world (NetSuite.com, 2024). It is a risk-based statute, built on the foundations of European values and ethics, seeking to position the region to be a global leader in AI innovation without harming the fundamental rights and well-being of Union citizens. Several countries are now looking to it as a possible blueprint for their own AI regulatory policies, including the Philippines.
In this post, we look back at the evolution of the EU AI Act, and examine its parallels and contradictions with a few of the AI regulatory proposals filed in the Philippines House of Representatives and its applicability in the broader Philippine context.
EU AI Act: From Conception to Enactment
The EU AI Act didn’t come into fruition overnight. It was a result of almost a decade of evolution that is grounded on Europe’s values- and rights-first culture and regulatory approach. (Bradford, 2023)
Before AI automation came into the public consciousness, ethical standards on robotics automation first took center stage in Europe, led by British Standard BS 8611 in 2016. In the same year, the IEEE Standards Association launched a global initiative on the Ethics of Autonomous and Intelligent Systems, with the purpose of drafting standards that will steer these technologies into becoming human-centered rather than an exclusively profit-driven endeavors, and effectively generalizing system autonomy to include intelligent systems beyond robotics. (Directorate General for Parliamentary Research Services, European Parliament, 2020)
On May 25, 2018, GDPR was enacted, setting the precedent for other successive technology legislation, including the AI Act, as it established that the EU is capable of regulating fast-changing digital technologies. It also extended the “Brussels Effect”, which recognizes the heavy influence of EU policies on global regulatory efforts owing to its unique geopolitical makeup (Bradford, 2012), to the digital space.
One of the major milestones is the Communication from the Commission to the European Parliament in 2018 which set the direction of where the AI legal and regulatory conversations should be heading. It argued that AI as a driver of economic growth should not be exclusive to maintaining alignment with European values, rights, and ethical standards.
In April 2019, The Ethics Guidelines for Trustworthy AI was published by the High-Level Expert Group on AI, convened by the European Commission. This established the common vocabulary for what “trustworthy AI” means, and three components that it should meet throughout its life cycle: lawful, ethical, and robust. These components were further broken down into seven core principles: human agency and oversight; technical robustness and safety; privacy and data governance; transparency; diversity, non-discrimination, and fairness; societal and environmental well-being; and accountability. It invited stakeholders to voluntarily integrate its proposed AI ethics assessment list into their systems.
The European Commission published The White Paper on AI in 2020 sounding the alarm on the dangers of a fragmented approach of EU member states to AI and pushed for a harmonized policy. This was also the beginnings of the risk-based approach that the EU AI Act is now well-known for, and proposed for stricter guidelines and obligations, while balancing excellence (innovation) with trust (rights preservation through AI regulation).
The first draft proposal for a concrete regulatory framework on AI was finally submitted by the Commission on April 21, 2021. This was groundbreaking, as it was the first ever comprehensive AI legal regulatory document in the world. It was cross-sector, and integrated the risk-based framework that the 2020 White Paper on AI suggested. It also framed AI as a product, which will thereby enable applicability of already existing mechanisms for product market surveillance and regulations.
What followed is a series of dialogues between the EU Council and Parliament. The former crafted their position in November 2022, with the latter releasing theirs in June 2023. A trilogue together with the EU Commission commenced soon after. The major contentions are listed in the table below.
The trilogues happened around the time LLMs like ChatGPT began to gain public adoption (OpenAI, 2022). It forced the policy makers to immediately include General Purpose AI (GPAI and Generative AI (GenAI) into the drafts for the act, with the Parliament’s position document explicitly mentioning ChatGPT in the text. A provisional agreement was finally reached in September 2023 after a “3-day marathon talks” (European Council, 2023), adopting the following changes:
Rules on high risk and high impact GPAI
A new governance architecture which would consist of an AI Office under the EU Commission, a panel of independent experts that would serve as GPAI advisors to the AI office, an AI Board under the EU Council, and an advisory forum of stakeholders
Prohibition list expansion, with exemptions for remote biometric identification by law enforcement in strict situations with safeguards
Pre-deployment impact assessment obligations of high-risk AI systems deployers
The AI Act was signed into law on June 13, 2024, published in the official journal a month after, and enacted on August 1, 2024. with its implementation to be conducted in phases. This sets the stage for the whole world to follow suit, as the Brussels effect is set in motion.
The EU AI Act as Benchmark for Philippine AI Regulation
The Philippines is racing to adopt its own AI regulatory legislation. In fact, 19 bills have been filed in the Congress just in the second half this year alone, not only reflecting the pressing need for a national AI policy, but also of the fragmented nature of local politics in the country. Whereas the EU brought forth its AI Act through collective effort, partly due to fears of fragmentation, the Philippine approach is grounded in individualism, personality politics, and optics (Leon, 2025). As evidenced by the surnames of some of the bill sponsors, powerful and influential political dynasties (Panti, 2025) are once again attempting to embed themselves in what could be one of the landmark legislations in the country. Political agenda aside, though, the large number of proposals signals that the Lower House is in the midst of recognizing and addressing the urgency posed by the lack of protections against potential misuse and abuse of AI.
The 19 house bills are presented in the succeeding table. For the rest of the section, we take on the challenge of consolidating the provisions in all of the bills, and attempt to perform a comparative analysis with the EU AI Act to determine the extent of its influence, and to spot stipulations that are unique to the Philippine context. Note that HB0013, HB00658, HB01920, HB02827, HB 03195, and HB05158 are all mirrors of each other, i.e. exactly the same bills but with different sponsors. Meanwhile, HB00073, HB00252, and HB03688 share the same title but only HB00073 and HB03688 are mirror bills. This means there’s effectively just 13 house bills, fewer albeit still considerable. To introduce a bit of clarity, the mirror bills are marked accordingly in the table. Additionally, for the rest of the text, we will only refer to the first bill (highlighted in boldface) to represent the group, and disregard its mirror filings.
Core Philosophies
The EU AI Act’s goal is to harmonize rules across member states while protecting fundamental rights. It relies on already existing legislation to provide mechanisms to protect human rights and welfare, while imposing the following obligations on providers and deployers of high risk AI systems to ensure safety, fairness, and accountability:
Risk Management Systems
Data Governance
Transparency and Human Oversight
Robustness, Accuracy, and Cybersecurity
On the other hand, a couple of Philippine proposals, HB00013 and HB00057, are aiming to leverage AI for national development while adopting rights protections through a more explicit “Bill of Rights”, seeking to empower citizens with the power to demand protection of their rights and to seek legal remedies whenever they are violated:
HB00013 Bill of Rights
Right to Protection from Unsafe and Ineffective AI Systems
Right Against Algorithmic Discrimination
Right to Privacy (compliance with Republic Act No. 10173, “Data Privacy Act of 2012”)
Right to Know
Right to Remedy
HB00057 Bill of Rights
Right to Human Agency
Right to Algorithmic Transparency
Right to Protection from Unjust Harm
Right to Data Sovereignty
Right to Safe and Accountable Systems
Right to Contestability and Redress
Right to Non-Obselescense
Right to Participate in Governance
Risk-Based AI Classification
A couple of the legal proposals in the Philippines for an AI regulatory framework seek to adopt a risk-based AI system classification, evidently influenced by the EU AI Act.
Although HB06028 only explicitly stated High, Moderate, and Low Risk, its risk tiers align with the High, Limited, and Minimal tiers of the EU AI Act. HB01196, meanwhile, is almost an exact replica of its EU counterpart.
AI Prohibitions
Zooming into the AI Prohibitions allow us to view key differences between the approaches, which is summarized in the succeeding table.
There is significant overlap on most of the harm categories. Glaring differences are the explicit inclusion of the ban of lethal autonomous weapon systems (LAWS) as emphasized multiple times in HB00013, and a strict stance on deep fake and misleading technologies in HB00057, HB00252, HB03480, HB03905, and HB06028 as opposed to softer transparency mandate in the EU AI Act.
Governance and Enforcement
The EU AI Act adopts a multilayered governance architecture by creating multiple regulatory bodies already mentioned earlier: the EU AI Office, European AI Board, Board of Scientific Experts, and Competent National-Level Authorities.
Meanwhile, several of the Philippine bills propose the adoption of a multi-agency approach, leveraging already existing government organizations through cross-functional collaboration. The proposed structure will include the following:
Philippine Council of Artificial Intelligence (PCAI) - multi-sectoral experts and advisers in charge of high level policy creation and will serve under the Department of Information and Communications Technology (DICT)
Artificial Intelligence Board (AIB) - primary regulatory and supervisory body composed of high-ranking government officials from different government organizations such as the DICT, Department of Science and Technology (DOST), Department of Trade and Industry (DTI), Intellectual Property Office of the Philippines (IPOPHIL), National Privacy Commission (NPC), and National Intelligence Coordinating Agency (NICA)
Artificial Intelligence Development Authority (AIDA) - alternative to AIDA, under the Department of Science and Technology (DOST)
National Center for AI Research (NCAIR) - primary body that will be tasked with advancing AI innovation and developing AI talent in the Philippines, and will be placed under either DOST or DICT
Penalties
The EU AI Act implements a tiered system of hefty fines:
Prohibited AI Practices - up to €35 million or 7% of total worldwide annual turnover
Non-Compliance - up to €15 million or 3% of global turnover
Inaccurate Information - up to €7.5 million or 1% of global turnover
The Philippine House Bills combine both fines and criminal liability that may lead to imprisonment:
Fines - up to ₱ 10,000,000
Imprisonment - from six months up to six years
Startups and SME Support
Both the EU and the Philippines recognize the need to allow small businesses and startups to flourish, and have included provisions that intend to provide space and flexibility for AI startups and SMEs in the form of regulatory sandboxes.
Regular fines in the EU require the business to pay the higher of the two options (hard cap or percentage cap) but small businesses may pay the lower of the two instead.
Some of the Philippine bills are seeking to provide funding and incentives to encourage building of AI businesses and ventures.
Protections Against Job Displacements
The EU AI regulation delegates worker protections to existing laws on social policy and labor. The Philippine Bills, in contrast, explicitly mandate a Job Displacement Program to be managed by the Department of Labor and Employment (DOLE) for job losses due to AI. Both require the disclosure of usage of AI to employees especially in high risk applications, and the need of human oversight to minimize risks. Employee training and reskilling on AI usage and development are also advised.
Summary
The EU AI Act extends on the legacy of the GDPR in setting the gold standard of technology legislation. It introduced a novel risk-based approach to AI governance and human rights protections. From surface level, it may seem like the EU AI Act is very lax in its enforcement of rights protections, e.g. no provisions for imprisonment, one must also consider that the law wasn’t designed to replace any criminal law that’s already existing (European Commission, 2021), and it actually relies on them to provide the legal framework for any violations that endanger human life and well-being.
While there are obvious inspirations taken by some of the Philippine bills from the EU AI Act, there are also efforts to differentiate the Philippine context from the EU experience. The political nature of the Philippines has resulted in a fragmented approach to drafting a national AI governance policy, making this particular comparative study extremely challenging to accomplish even with the presence of LLMs that aid in making sense of all the documents. I do recognize, though, that had it not been for LLMs, this would not have been possible. It is also worth noting that this paper’s scope was confined to bills filed in the Philippine House of Representatives, and deliberately left out similar initiatives in the Philippine Senate.
The Philippines is still in its early stages of laying down the groundwork for a national AI law. Hearings on these bills are ongoing as I write this, and they will need to overcome both legal and political hurdles before they can all be consolidated into a unified initiative. The president has not even included the need for an AI law in his list of priority bills, choosing to shift the government’s focus instead to anti-corruption measures, and for good reason (Mangaluz, 2025). As a citizen and technology professional myself, I can only hope that the Philippine AI Act that finally comes out of this will be one that serves the interests of the people, not of its politicians nor of the huge conglomerates that dominate Philippine society, and that its enforcement will not go the route of the many laws in the country: poor to no enforcement at all (Philipp, 2023; The Philippine Star, 2021).
Acknowledgement
I would like to express my gratitude to AI ethicist and friend Doc Ligot, for making these bills available publicly through his personal Google Drive. These legal documents would most likely not have been accessible without his efforts.
AI Usage Disclosure
In order to augment my lack of legal training and skills, I utilized LLM technologies. ChatGPT was used in order to gain better understanding about the evolution of the EU AI Act. Meanwhile, to consolidate and synthesize the 19 bills currently filed in the Philippine Congress, and conduct a comparative analysis with the EU AI Act, I turned to Google’s NotebookLM.
However, while content in this paper is the result of interactions with ChatGPT and NotebookLM and may have been inspired by them, none of the material has been directly generated by AI. Additionally, I had to perform due diligence in making sure that the output is grounded on the source materials, and not from AI hallucination.
Sources
An Act Establishing a Comprehensive Legal Framework for the Development, Deployment, and Regulation of Artificial Intelligence in the Philippines, Creating the Artificial Intelligence Development Authority (AIDA), and for Other Purposes, No. 1196, House of Representatives (2025).
An Act Establishing a Legal Framework for the Responsible and Accountable Use of Artificial Intelligence in the Philippines, No. 3480, House of Representatives (2025).
An Act Establishing a Regulatory Framework for a Robust, Reliable, and Trustworthy Development, Application, and Use of Artificial Intelligence Systems, Creating the Philippine Council on Artificial, Delineating the Roles of Various Government Agencies, Defining and Penalizing Certain Prohibited Acts, 13, House of Representatives 1 (2025).
An Act Instituting the National Artificial Intelligence Code of the Philippines, Creating the Bureau of Artificial Intelligence Systems Under the Department of Information and Communications Technology, and Establishing a Comprehensive Framework for the Safe Ethical and Inclusive Governance of Artificial Intelligence Systems in the Country, and for Other Purposes, 57, House of Representatives 1 (2025).
An Act Mainstreaming Artificial Intelligence Into Government Policy Formulations, Establishing the Framework Strategy and Program on Artificial Intelligence, Creating for This Purpose the Artificial Intelligence Commission, and for Other Purposes, No. 2323 (2025).
An Act Mandating the Department of Education, Commission of Higher Education, Department of Science and Technology and Department of Information and Communications Technology to Develop Generative Artificial Intelligence (GenAI) Regulations to Maximize Its Positive Impact on Philippine Education System, 598, House of Representatives 1 (2025).
An Act Promoting Data-Driven Governance and Artificial Intelligence Solutions in the Public Sector Through Government-Industry-Academia Collaboration, and Appropriating Funds Therefor, 1484, House of Representatives 1 (2025).
An Act Promoting the Development and Regulation of Artificial Intelligence in the Philippines, 73, House of Representatives 1 (2025).
An Act Promoting the Development and Regulation of Artificial Intelligence in the Philippines, 252, House of Representatives 1 (2025).
An Act Promoting the Development and Regulation of Artificial Intelligence in the Philippines, Creating for the Purpose the Philippine Artificial Intelligence Commission, and for Other Purposes, No. 3462, House of Representatives (2025).
An Act Providing a National Strategy for the Development of Artificial Intelligence, Creating for the Purpose and National Center for Artificial Intelligence Research, and Appropriating Funds Therefor, 659, House of Representatives 1 (2025).
An Act Recognizing the Use of Artificial Intelligence and Generative Technologies in the Commission of a Crime as an Aggravating Circumstance, and Providing for the Liability of Developers and Platforms Thereof, 3905, House of Representatives 1 (2025).
An Act Regulating the Development and Use of Artificial Intelligence Systems in the Philippines, Promoting Ethical and Responsible Artificial Intelligence Innovation, and Integrating Sustainability and Futures Thinking in National Policy Making, and for Other Purposes, 6028, House of Representatives 1 (2025).
Bradford, A. (2012). The Brussels Effect. In A. Bradford, The Brussels Effect (1st ed., pp. 25–66). Oxford University PressNew York. https://doi.org/10.1093/oso/9780190088583.003.0003
Bradford, A. (2023). The European Rights-Driven Regulatory Model. In A. Bradford (Ed.), Digital Empires: The Global Battle to Regulate Technology (p. 0). Oxford University Press. https://doi.org/10.1093/oso/9780197649268.003.0004
Communication From the Commission to the European Parliament, the European Council, the Council, the European Economic and Social Committee and the Committee of the Regions - Artificial Intelligence for Europe (2018). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52018DC0237
Directorate General for Parliamentary Research Services, European Parliament. (2020). The ethics of artificial intelligence: Issues and initiatives. Publications Office. https://data.europa.eu/doi/10.2861/6644
European Council. (2022, November 25). Proposal for a Regulation of the European Parliament and of the Council laying down harmonised rules on artificial intelligence (Artificial Intelligence Act) and amending certain Union legislative acts—General approach. Consilium. https://data.consilium.europa.eu/doc/document/ST-14954-2022-INIT/en/pdf
European Council. (2023). Artificial Intelligence Act: Council and Parliament strike a deal on the first rules for AI in the world.
European Parliament. (2023). Parliament’s negotiating position on the artificial intelligence act.
High-Level Expert Group on AI. (2019, April 8). Ethics guidelines for trustworthy AI | Shaping Europe’s digital future. https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai
Leon, D. de. (2025, June 30). Your lawmaker filed either hundreds, or just a few bills. So what? RAPPLER. https://www.rappler.com/newsbreak/explainers/context-nuances-how-many-bills-resolutions-filed-congressman/
Mangaluz, J. (2025, December 9). Marcos to Congress: Prioritize Anti-dynasty bill, other transparency measures. Philstar.Com. https://www.philstar.com/headlines/2025/12/09/2493003/marcos-congress-prioritize-anti-dynasty-bill-other-transparency-measures
NetSuite.com. (2024, December 4). What Is General Data Protection Regulation (GDPR)? Oracle NetSuite. https://www.netsuite.com/portal/resource/articles/business-strategy/gdpr.shtml
OpenAI. (2022, November 30). Introducing ChatGPT. https://openai.com/index/chatgpt/
Panti, L. (2025, December 10). Anti-political dynasty bill: Ever present and absent at the same time. GMA News Online. https://www.gmanetwork.com/news/topstories/nation/969040/anti-political-dynasty-bill-ph-timeline/story/
Philipp, J. (2023, April 3). Strengthening the Rule of Law in the Philippines. The Borgen Project. https://borgenproject.org/rule-of-law-in-the-philippines/
Proposal for a Regulation of the European Parliament and of the Council Laying Down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act) and Amending Certain Union Legislative Acts (2021). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52021PC0206
Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the Protection of Natural Persons with Regard to the Processing of Personal Data and on the Free Movement of Such Data, and Repealing Directive 95/46/EC (General Data Protection Regulation) (Text with EEA Relevance), 119 OJ L (2016). http://data.europa.eu/eli/reg/2016/679/oj
Regulation (EU) 2024/1689: Harmonised Rules on Artificial Intelligence (2024). http://data.europa.eu/eli/reg/2024/1689/oj
The Philippine Star. (2021, October 17). EDITORIAL - Weak rule of law. Philstar.Com. https://www.philstar.com/opinion/2021/10/17/2134647/editorial-weak-rule-law
White Paper on Artificial Intelligence - A European Approach to Excellence and Trust (2020). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52020DC0065