We’ve taken a look at bills on social media regulation, deepfakes, and disinformation. Now, we look at another important technology issue, artificial intelligence.

Its emergence into the mainstream began a few years ago when ChatGPT became a hit, and saw furious adoption after people saw how far the technology has come. 

In the 20th Congress, the technology has been a focal point as well, with about 40 House bills and 7 Senate bills, all looking to explore possible guardrails, promote AI education, enforce labor protections, establish councils and committees and more. 

Here’s the current lay of the land. 

National agencies for AI

Several bills propose the establishment of dedicated authorities to serve as a sort of central nervous system for AI governance, each with distinct mandates for regulation, administration, and enforcement. 

The Bureau of Artificial Intelligence Systems (BAIS), proposed under several measures such as HB 57, 7627, and 9304, is designed to “serve as the national regulatory, technical, investigatory, and administrative body on matters relating to the governance of Artificial Intelligence systems in the Philippines.”

Alternatively, HB 73, 1196, and 3688, alongside SB 1190, propose an Artificial Intelligence Development Authority (AIDA), which would be “created and established to oversee the development and deployment of AI technologies, ensuring compliance with AI ethics principles and guidelines, and protecting the rights and welfare of individuals and communities affected by AI technologies.”

Other measures suggest the Philippine Council on Artificial Intelligence (PCAI), envisioned as a “policy-making and advisory body of experts under the Department of Information and Communications Technology (DICT).”

Complementary to these oversight structures is a proposed National Registry of AI Systems, said to be a mandatory database requiring developers and deployers to register their technologies before public use. 

This registry is intended to ensure transparency, especially for so-called high-risk systems, by documenting their intended purpose, data sources, and built-in safeguards. 

Furthermore, HB 3905, the AI-Aided Criminality Liability Act, specifically focuses on platform and developer accountability, holding corporations and organizations civilly or criminally liable if they “knowingly or negligently enable the use of their AI technologies or infrastructures for the commission of crimes.”

These frameworks aim to achieve a more unified enforcement mechanism for the rapidly evolving AI landscape.

An AI Bill of Rights

Congress also seeks to institutionalize an AI Bill of Rights. Found across multiple measures, this bill of rights affirms that individuals interacting with AI must retain their “digital dignity,” meaning they are treated as humans with agency rather than merely as data points. 

Key protections include the Right to Meaningful Human Oversight, which mandates that no AI system has final authority over decisions affecting an individual’s rights or safety without a mechanism for human review and override. 

Other essential rights include the Right to Transparency and Awareness, ensuring individuals know when they are subject to AI-driven decisions; the Right to Privacy and Data Sovereignty, which aligns AI use with the Data Privacy Act of 2012; and the Right to Redress, providing clear avenues for appeal when AI produces harmful or erroneous outputs.

The proposed regulatory philosophy is complemented with a risk-based classification system which determines the level of government scrutiny based on an AI’s potential impact. 

High-risk systems are those deployed in critical sectors such as healthcare, finance, law enforcement, and critical infrastructure, where errors could materially affect fundamental rights or public safety, as defined in bills such as SB 2097 and HB 6028.

These systems face the strictest requirements, including mandatory assessments for the impact of their algorithms, bias testing, and continuous auditing. 

Conversely, moderate-risk systems with limited impact undergo lighter oversight, while low-risk systems, such as basic data visualization tools, operate with minimal regulatory intervention. 

Certain practices are categorized as unacceptable risks (HB 1196) and are strictly prohibited, including the use of AI for generalized social scoring (HB 6423), unauthorized mass biometric surveillance, and manipulative technologies designed to distort human behavior beyond a person’s consciousness.

Labor protections 

Legislators appear to be deeply concerned about the potential for AI to displace workers and undermine democratic processes, leading to many protective measures in the labor and electoral sectors. 

The labor-focused bills — HB 387, 1899, 3908, 3921, and 4433, as well as SB 182 — seek to strike a balance between harnessing the productivity of AI and protecting livelihoods. 

These bills attempt to mandate that employers cannot terminate employees based on the availability of AI tools, requiring instead meaningful human review and a “bona fide” justification for the introduction of automation. 

Proposed protections include mandatory reskilling and upskilling programs to help workers transition into new roles.

For elections, HB 3691, 5696, 6313, 6623, 7326, and 8734 focus on monitoring AI during electoral periods. These measures respond to the threat of disinformation and political manipulation by requiring candidates and political parties to register their digital campaign platforms with the Commission on Elections (Comelec). The legislation mandates clear and conspicuous disclosure for all AI-generated campaign materials, including watermarks and detailed explanations of the nature of the manipulation. 

Additionally, the bills prohibit the use of “false amplifiers” like bots and the dissemination of deceptive deepfakes intended to manipulate voter behavior, granting Comelec the power to issue urgent takedown orders for high-harm content.

Education and child protection

Several bills focus on building AI capacity through education and innovation. 

Measures like HB 243, 598, 2186, 6172, and 7738 seek to integrate AI education into the national curriculum from elementary to tertiary levels. The impetus is to ensure that future graduates are not merely passive users of technology but innovative contributors who understand algorithms, data privacy, and AI ethics. 

Proposed plans include creating age-appropriate modules for K-12, introducing AI-focused degree programs in universities, and providing teacher capacity-building to ensure effective instruction. 

National innovation is further supported through plans for National AI Infrastructure, which would provide shared access to public computing facilities, and open-source AI libraries. 

To foster local talent, many bills also mention “regulatory sandboxes” which are supervised environments for testing experimental AI systems with time-bound exemptions from certain compliance requirements. 

Finally, responding to the “terrifying speed” of the evolution of the digital landscape, HB 2319 and 7858 focus on protecting minors by amending the Anti-Online Sexual Abuse or Exploitation of Children (OSAEC) Act. These amendments specifically criminalize the creation, possession, and distribution of child sexual abuse materials (CSAEM) generated through AI, ensuring that the law addresses AI-specific cybercrimes that exploit children. – Rappler.com