The word “democracy” comes from two Greek words: demos, meaning people, and kratos, meaning power. Together, they mean “power of the people.”
A similar idea has emerged in the world of artificial intelligence.
IBM defines AI democratisation as the more equitable spread of AI applications and capabilities across society. In practice, it means making AI accessible not only to developers, engineers, and large technology companies but also to businesses, governments, communities, and individuals.
As AI increasingly embeds in daily life, industry leaders argue that democratising AI means sharing the power to use, develop, and govern the technology, rather than being concentrated in the hands of a few.
For executives at Red Hat, democratisation begins with one fundamental principle: choice.
The power of choice
Daniel Aw, VP and GM of Asia Pacific and Japan at Red Hat, believes AI is at a crossroads similar to the one the computing industry faced three decades ago.
“I think when we talk about democratising AI, it’s that you do not want to be locked into a certain vertical stack,” he said.
Aw compared today’s AI landscape to the Unix era, when organisations were often dependent on one ecosystem, limiting flexibility and innovation.
“AI is the same way,” he said. “The AI moment today is similar to the Linux moment many years ago.”
Instead of being tied to a single vendor, organisations should have the freedom to choose the hardware, software, and models that best fit their needs.
Challenges unique to Asia-Pacific
While enthusiasm for AI continues to grow, democratisation faces several challenges across the Asia-Pacific.
According to Vincent Caldeira, chief technology officer for Asia-Pacific at Red Hat, one of the region’s biggest hurdles is infrastructure.
“First, we have to catch up on AI factories, so making GPUs available,” he said.
Many governments across the Asia-Pacific are already investing in sovereign AI initiatives and local AI infrastructure. Without sufficient computing power, however, organisations cannot fully participate in the AI economy.
Beyond infrastructure, Caldeira believes cultural diversity presents both a challenge and an opportunity.
Asia-Pacific is home to thousands of languages, cultures, and ways of thinking. Building AI systems that merely understand local languages is not enough.
“There is a big difference between tuning an AI model with data that is culturally from your country and just the language,” he said.
“You’ve got such a rich diversity of thinking,” Caldeira said. “If AI technology is in the hands of very few conglomerates, you are going to lose that diversity and that culture.”
If AI technology is in the hands of very few conglomerates, you are going to lose that diversity and that culture. Vincent Caldiera
Prem Pavan, vice president and head of the APAC Partner Ecosystem at Red Hat, highlighted another concern: uncertainty.
“The challenge is about the future and the unknowns that the future has in place,” he said.
Organisations must make technology decisions today without knowing how quickly AI will evolve tomorrow. Those choices will determine their ability to adapt to new technologies, changing regulations, and emerging business needs.
Sovereignty is also becoming increasingly important.
“What are the choices that I am making today that will define my ability and flexibility tomorrow?” Pavan asked.
For many enterprises and governments, the question is not only how to adopt AI, but how to maintain control over their data, infrastructure, and long-term technology strategy.
What are the choices that I am making today that will define my ability and flexibility tomorrow? Prem Pavan
Open source as an equaliser
For Red Hat executives, open-source plays a central role in making AI more accessible.
Pavan described open source as providing both the freedom of choice and the freedom to change course when needed.
“It is the power of choice and the flexibility of choice that you have in the context of democratisation,” he said. “It is also the flexibility of exit.”
Organisations may need to adapt because technology evolves or because geopolitical and sovereignty considerations change. Building on open-source foundations can provide the flexibility to respond to those shifts.
Caldeira sees open source as a way to expand access to AI beyond a small group of highly specialised experts.
Several years ago, developing AI often required teams of highly trained data scientists who understood everything from machine learning algorithms to infrastructure optimisation.
Today, the goal is broader participation.
“We want a much broader population of people to access and consume it,” he said.
Rather than requiring every user to understand highly technical concepts and jargon, open platforms simplify the complexity.
At the same time, open-source communities can collectively build technologies that allow organisations to operate AI systems on their own infrastructure.
“No single company in Indonesia or India can compete with an OpenAI, Alibaba, or Google,” Caldeira said. “However, if the global community together contributes open-source capability into projects, we basically have a very powerful stack that can be used by many.”
For Caldeira, the growing involvement of major industry players in open-source initiatives is evidence of the model’s success.
“NVIDIA is one of the richest companies in the world,” he said. “Yet NVIDIA still chooses to work with the open-source community.”
Additionally, open-source technologies, in Aw’s view, help prevent that lock-in.
While open-source AI models may not always match the scale of frontier models developed by major technology companies, they provide what many organisations with flexibility, transparency, and innovation.
“Open source allows the ecosystem to grow, not just the company that’s providing the vertical stack,” Aw said.
Innovation in the AI era is increasingly being driven by communities rather than individual corporations. Aw pointed to projects that began as personal initiatives and quickly gained momentum through open-source collaboration.
“We always have this saying: you stand on the shoulders of those who came before you,” he said. “You don’t need to start from zero.”
While open-source models offer flexibility and transparency, enterprises must still consider support, security governance, and operational complexity.
“You stand on the shoulders of those who came before you. You don’t need to start from zero,” Daniel Aw
The future is collaborative
Looking ahead, Red Hat executives envision a future where countries and organisations collaborate rather than build AI capabilities in isolation.
Caldeira suggested that nations could jointly develop foundational AI components instead of independently investing in similar infrastructure and models.
“What if those countries would come together and contribute on a joint solution?” he asked.
Shared investments in infrastructure, data pipelines, and model development could reduce costs, accelerate innovation, and expand access to AI capabilities across the region.
Ultimately, democratising AI is about ensuring that technological progress benefits society as a whole.
In a region as diverse as the Asia-Pacific, openness, collaboration, and flexibility are its strengths.
AI should not belong only to developers, engineers, or multinational corporations. Democratisation means giving more people the opportunity to participate in shaping the technology and its future.
If democracy is the power of the people, then democratising AI may be about ensuring that power remains in the hands of many rather than a select few.
