BUDAYA AI: A Malaysian vision for inclusive AI governance that can inspire global AI development

Chithra, Michael, and I met working for Malaysia’s National AI Office (NAIO) earlier in the year. The three of us were members of their AI Governance Working Group—a pool of experts pulled together to advise the Malaysian government on AI-policy more broadly.

Working with NAIO and the other fine folks in the working group, it dawned on us that Malaysia had a golden opportunity to try out something truly novel in the AI governance space. In the subsequent piece we wrote for The Edge Malaysia, we set out our vision for what that is: BUDAYA AI.

Tech billionaires, half a world away, evangelise an AI future that conveniently benefits them the most. Their “innovations” tend substantially to concentrate economic and, as a result, political power in their hands. Countries in the Global Majority are compelled to get on board with their vision and adhere to their rules, or are admonished for falling behind.

The technologies they export carry embedded biases and assumptions from their societies. These tools often reflect centuries of inequality shaped by profit motives, representing a new form of technocolonialism that extends digital domination across borders. Imperfect systems are “shipped” to market quickly and any so-called bugs are (sometimes) iterated out in future versions, after the products have been tested on “users”. But these aren’t bugs, they are harmful to people, communities and systems of democratic governance.

Countries in the Global Majority are not passive beneficiaries of some inevitable trajectory of technology. We have the power to bend the arc of technological innovation in ways that we want for our societies. We possess the agency and responsibility to shape innovation through public policy. By centering human dignity and social cohesion, we can chart alternative courses so that technology serves society, not the other way around.

AI must be designed and deployed in ways that reflect and protect the world’s rich natural and multicultural identity. In a society shaped by diverse languages, faiths, traditions and norms, technology must support — not flatten — cultural expression.

Malaysia's multicultural democracy, therefore, provides an alternative to predominately Western approaches to AI governance. Its multicultural heritage presents a collective wisdom that strengthens our agency as we shape the arc of technology. To this end, Malaysia's National AI Office (NAIO) asked us to explore AI governance principles that align with Rukun Negara and Prime Minister Datuk Seri Anwar Ibrahim's Madani framework. As we worked to combine Malaysian insights with international AI governance expertise, we uncovered something far more insightful than anticipated. We realised that Malaysia's diversity fuels a moral compass rooted in constitutional principles and lived multicultural experiences and values, which are not subordinate to the market. Too often, the guiding principles handed down through international bodies become siloed and fail to articulate the end state or the overall social outcomes we want to see. Too often these put “the market” on a pedestal such that core societal values are subordinated. Malaysia’s moral compass, therefore, represents an unprecedented opportunity for global leadership. It grants Malaysia the chance to pioneer the world's first AI governance model grounded in multicultural democracy and the civic values that guide society. Society’s core values are articulated in the Madani framework, which encompasses sustainability, prosperity, innovation, respect, trust and compassion. It provides an excellent foundation upon which to create policies directly linked to the social outcomes we want from AI systems. Malaysia's multicultural reality offers something few other nations possess: decades of successful harmony in diversity.

As a global AI governance crisis unfolds before us, unmitigated harms challenge the social outcomes we desire. Facial recognition systems work effectively for white, male faces but fail for women and people of colour. Hiring algorithms systematically discriminate against women and minorities in workplace decisions. Workplace surveillance systems disempower workers and erode their bargaining power. Credit scoring systems reinforce historical inequalities rather than addressing systemic problems. Individual agency diminishes as AI systems make opaque decisions for people that impact their social protection, access to financial means, even jobs.

AI systems, like large language models (LLMs), crowd out minority voices and perspectives in mainstream discourse. In Malaysia's context, this manifests in concrete ways when customer service chatbots cannot comprehend Kelantanese Malay, voice assistants fail to recognise Sabahan creole speakers, translation systems homogenise regional expressions into standardised forms, and AI tutoring systems cannot adapt to indigenous learning patterns. Systems trained on formal Bahasa Malaysia systematically erase dialects. LLMs trained solely on formal texts risk eliminating the rich linguistic diversity that defines Malaysian identity. And the global dominance of English and Anglo-centric behaviours, priorities, and analytic approaches, etc., weakens our cultures and languages. These are not glitches but features of current innovation practices. They arise when homogeneous teams design AI based on their own experiences. Scant opportunity for meaningful inclusion of diverse voices (literally and figuratively). The development process seldom, if ever, involves genuine consultation with affected communities, which may be half a world away.

Silicon Valley grapples with algorithmic bias and the concentration of power in the hands of private companies and their billionaire owners. Beijing exploits minorities to sharpen surveillance tools and sees frontier tech as a means for wielding more control over citizens. India's embrace of its Aadhaar biometric system marginalises minority communities from essential services. The European Union struggles to achieve consensus across its 27 member states, facing America’s unrestrained neoliberal politics and the pressure of big tech bearing down upon it. The rest of the Global Majority often appears as window dressing, participating in global agenda-setting moments without real influence or power.

Malaysia stands in a novel position to leverage budaya, or culture, as a defining feature of its approach to governance of the global AI ecosystem.

Inspired by Malaysia’s experience, the BUDAYA AI framework illustrates how social outcomes based on core civic and moral values can be achieved with AI development guided by effective policy. In other words, these values serve as a moral compass for public policy, grounded in Malaysia’s rich heritage.

The Rukun Negara has maintained peaceful coexistence since independence, facilitating growth in our multi-ethnic, multi-faith society for over six decades. Contemporary articulations of national purpose and ethical governance, such as the Madani principles, provide a holistic understanding of what we as a society aim to achieve. Madani, for example, reminds us that progress must be measured not only in GDP growth or technological capability or increases in shareholder value, but in the extent to which innovations uphold justice, dignity, and opportunity for all.

For example, kesejahteraan (prosperity/holistic well-being) is not just the absence of suffering or poverty, but the presence of justice, dignity, harmony, and opportunity for all. It’s a moral and civic ideal that guides Malaysia towards becoming a compassionate, inclusive, and thriving society rooted in ethical leadership and shared prosperity. It articulates both principles and outcomes. The other Madani principles function similarly to ensure that all citizens have the opportunity to thrive in a just and compassionate society:

• Kemampanan (Sustainability) ensures long-term cultural preservation through inclusive AI that protects rather than erases diversity.

• Daya Cipta (Innovation) propels culturally-informed technological advancement that honours tradition while embracing progress.

• Hormat (Respect) ensures recognition of all community voices in AI decision-making processes.

• Keyakinan (Trust) fosters transparent AI systems that are accountable to the people they serve.

• Ihsan (Compassion) places human dignity above profit margins in every algorithmic decision.

Collectively, these give rise to indigenous expressions of moral and social wisdom that the legalese frameworks we are so used to can replicate.

We acknowledge the importance of other AI areas, including regulation, talent development, safety, sovereignty, security, among others. However, BUDAYA AI regards governance and ethics as primus inter pares (first among equals) among all of them, a first position among equals in AI policy prioritisation. Governance provides the moral lighthouse that guides policy decisions that align with societal outcomes our moral compass demands. Malaysia's AI development adheres to shared principles of social cohesion and civic responsibility. This highlights the guardrails that are commonly recognised and that emerge organically from our society.

Malaysia's success in maintaining peace among diverse communities offers a proven blueprint for AI governance. Our traditional dances tell stories of inclusive innovation. Zapin celebrates cultural exchange and adaptation. Sumazau expresses gratitude for abundance while respecting spiritual beliefs. Chinese dragon dances symbolise collective strength that overcomes adversity, and Indian classical forms like Bharata Natyam convey over 100 movement patterns reflecting life's complexity. Much as these cultural elements form the fabric of our lives and guide how we live with one another, they must also guide how technology (that is woven into our lives, often whether we want it to or not) develops. Not the market.

These cultural expressions embody some of the fundamental principles necessary for ethical AI. Just as traditional musicians blend gambus, sape, and tabla to create harmonious ensembles, BUDAYA AI forms the DNA of the algorithms that make up AI systems. The concept of "budi" (gratitude) combined with "daya" (action) creates "budaya" (culture). Its very etymology foregrounds human-centred AI use that emerges when technological capability is built on a strong foundation of dignity and respect.

BUDAYA AI operationalises governance founded upon clear civic values that lead to just and equitable societal outcomes conducive to human and environmental flourishing. It proposes concrete mechanisms that convert these values into functional systems. For example, community oversight networks empower marginalised voices in AI deployment decisions. Drawing from traditional kampung councils and longhouse decision-making, these networks possess veto power over AI deployments that affect their communities, with mandatory consultation periods that reflect customary consensus-building practices. Just as traditional gotong-royong (mutual cooperation) unites communities for collective benefit, these networks ensure that no community is overlooked during digital transformation.

Similarly, multi-faith ethics panels ensure that AI systems respect all religious perspectives equally. These panels comprise Islamic scholars, Buddhist monks, Christian pastors, Hindu priests, and indigenous spiritual leaders, ensuring that AI systems respect religious sensitivities in healthcare, education, and social services. They draw upon Malaysia's successful tradition of interfaith dialogue during festivals such as Deepavali, Chinese New Year, and Hari Raya.

The integration of indigenous knowledge and lived experiences transforms traditional wisdom into strategic assets. The ecological understanding of the Orang Asli guides sustainable AI practices that grasp environmental interconnections. Sabah's Kadazan wisdom regarding harvest cycles informs temporal AI decision-making that respects natural rhythms. The Peranakan baba-nyonya traditions of cultural fusion guide how AI systems can blend various approaches respectfully while maintaining their distinct identities, ensuring that no aspects of the multicultural diversity are crowded out by algorithms slowly nudging people away from them.

Linguistic preservation infrastructure safeguards over 140 languages in Malaysia from AI homogenisation. World-class research infrastructure effectively preserves and applies indigenous knowledge systems. The rich dialects from Kelantan to Sarawak, and beyond, inform natural language processing that truly understands communities’ communication patterns rather than imposing standardisation under the demands of so-called efficiency.

Traditional knowledge holders become partners, not subjects, in AI development. The philosophical concept of "budi" ensures that AI development remains appreciative of existing wisdom rather than replacing it with technologically deterministic solutions.

Malaysia’s diversity remains instrumental to its identity as a nation emerging from colonialism while also providing a strategic competitive advantage in the age of AI. An infinite human resource that perpetually creates value.

When perspectives from the country’s Malay, Chinese, Indian, and indigenous communities influence AI systems, we develop more robust and intelligent technology that effectively serves diverse populations. When urban and rural communities participate equitably in shaping our technological future, we avert digital divides that fragment society. When multiple languages and cultures inform training data, trade-offs and tech outcomes, AI becomes more meaningfully intelligent rather than merely sophisticated within narrow parameters.

Religious diversity offers ethical insights for algorithmic decisions that require moral reasoning beyond simple utilitarian calculations and efficiency trade-offs. The cultural complexity serves as a test for inclusive AI that must navigate conflicting values and interests. Malaysia can assess universal AI solutions in our diverse society, as if they succeed here, they can succeed elsewhere.

During his May 2025 Senate testimony, OpenAI chief executive officer Sam Altman advocated for global adoption. He argued that widespread American-developed technology enhances national security and international influence. This push for technological dominance exemplifies what critics refer to as algorithmic technocolonialism. Such approaches impose the digital frameworks of dominant nations on populations worldwide.

BUDAYA AI offers a principled counter-narrative through innovative jurisdictional approaches. It focuses on Malaysian cultural frameworks and values within AI governance systems. This initiative enhances digital sovereignty and shields citizens from external technological influence. It also defends against cultural appropriation through technological measures.

BUDAYA AI shows that effective AI governance does not have to follow a single standard. Instead, it suggests locally rooted approaches that develop stronger collective systems. This is more valuable than any single, unified framework could ever be. This approach offers blueprints for other nations aiming to safeguard their digital autonomy. Countries can engage in global AI ecosystems on their own terms.

This represents a pragmatic strategy based on market realities and competitive advantages within the emerging global AI economy. The world is much more diverse than the homogeneity that predominates among the tech’s powerful elite class that is currently admonishing the world to follow their technological vision of the future. The market for just and equitable AI systems represents limitless opportunities. Malaysia is exceptionally positioned to capitalise on this opportunity through its established expertise in multicultural governance, interfaith and intercultural community organisations and post-colonial nation-building.

AI systems operating across cultural boundaries will emerge as strategic assets in an interconnected world where businesses must cater to diverse populations. As more governments seek to guide the arc of technology as it impacts their citizens and societies, demand for such responsible AI systems, including governance models, will only increase. Malaysia thus offers not just technology but governance blueprints to operationalise cultural sensitivity and social inclusion at scale. This transforms cultural alignment from an abstract theory into exportable intellectual property that fosters sustainable competitive advantages.

When multinational companies require AI for diverse markets, they seek partners who comprehend diversity at fundamental levels, rather than merely surface cultural adaptations. When researchers examine inclusive AI governance, they will study BUDAYA AI as the model governance structure that aligns AI with desired social outcomes. When nations seek models for multicultural AI deployment, Malaysia can serve as both an inspiration and a technical partner.

While other nations struggle to balance diversity amid polarising politics while failing to articulate a holistic vision for a post-AI society that is just, equitable and thriving, Malaysia has successfully woven together society much like it weaves the most colourful and imaginative batiks. This operational experience in multicultural harmony and moral clarity over the kind of society it wants, directly translates to AI governance capabilities that few other nations have seriously developed. These market realities favour Malaysian leadership in inclusive, responsible AI development.

The Ministry of Digital faces a civilisational choice about Malaysia's future that will determine whether Malaysia leads or follows in the AI era. Will Malaysia chase Western-defined innovation that perpetuates technocolonialism, or will it pioneer culture-first AI ecosystems that respect human dignity? Profit-centric models extract value, while BUDAYA AI generates shared dignity across communities. Resisting this Silicon Valley Imperialism further presents Malaysia with legion opportunities to further shape the global AI governance discourse.

Malaysia can aim to establish a Global AI Cultural Summit in Kuala Lumpur even this year, creating BUDAYA AI certification standards, assessment tools, and process flows for inclusive AI systems that serve as international benchmarks, informing global ISO discussions. Malaysian universities can evolve into hubs for culturally aware AI research that attracts global talent and investment. Local startups can develop AI solutions that operate across diverse cultures, positioning themselves as strategic assets in the global marketplace. This is more than just a possible future for Malaysia, it is achievable as we position BUDAYA AI at the heart of inclusive global dialogue on how best to guide this frontier technology.

BUDAYA AI redistributes AI authority, and power, from technological elites to panels led by citizens, democratising digital sovereignty and placing power into the hands of the people instead of concentrated corporate interests. Malaysia can demonstrate that it is possible to build AI systems that benefit humanity without yielding to neoliberal economics that perpetuate inequality, concentration of power and ultimately impinge on citizens’ ability to flourish by building the lives they want.

Malaysia is well positioned to respond to the world's call for unity in an era of growing polarisation. Multicultural democracy provides the cultural competence to deliver inclusive technology that works for everyone. BUDAYA AI offers the governance framework necessary to realise this vision. Just as Malaysia's Petronas Towers symbolised the capabilities of an emerging nation in the physical realm, BUDAYA AI can establish itself as the global standard. Demonstrating that inclusive governance leads to superior technology, not despite cultural diversity, but because of it.

The world is watching, waiting, and desiring leadership that exemplifies unity in diversity rather than enforced homogenisation and dominance. Malaysia has both the moral authority rooted in constitutional democracy and the practical framework embedded in its values and moral framework as a nation. Our multicultural society, protected by law and guided by shared values, presents a unique model that other nations find difficult to replicate.

We believe Malaysia must choose to seize this moment, to lean into an inclusive AI future that unites rather than divides. BUDAYA AI offers a pathway to turn this into reality, a pathway found by bringing Malaysia's established multicultural wisdom as a key tool in both harnessing and managing the power of AI models and those who make them. This represents more than technology policy. This represents a civilisational choice between extractive technocolonialism and inclusive digital systems that place human dignity at the centre of governance.