Many businesses in the Philippines are exploring open-source artificial intelligence (AI) models from China, drawn by the promise of affordability and flexibility. But while these tools are often marketed as “free,” experts warn that the actual cost of adopting and using them can be much higher than expected.
“Open-source AI can be a good option for companies with the right resources, but calling it ‘free’ is misleading,” said Ferdie Saputil, country director for the Philippines at Searce. “You still need to invest in infrastructure, talent, and compliance. Without those, the model won’t work the way you expect.”
AI models like DeepSeek and Qwen are available for download at no cost, but using them properly requires significant computing power. Most small and medium-sized businesses don’t have the high-end GPUs or upgraded networks needed to run these models. Energy bills can also increase, adding to the financial burden.
Then comes the human challenge. AI tools need skilled people to set them up and keep them running. In Southeast Asia, finding experienced AI talent is difficult. Even when companies train their own staff, they risk losing them to better offers from larger firms.
“Even if you get the tech right, AI adoption fails without the right people,” Saputil said. “Upskilling your team is essential, but it takes time and comes with the risk of losing them to higher-paying companies.”
Regulatory compliance is another hidden hurdle. The AI field is constantly changing, and ASEAN countries each have their own rules. Using models from China may also raise concerns about data privacy, cross-border data transfers, and long-term legal risks. Companies might find themselves needing to rebuild or adjust their systems to stay within legal guidelines.
Talent and training
According to Saputil, the biggest obstacle in AI adoption isn’t the technology, it’s the people using it. Many companies underestimate the training and change management needed for success. Without proper onboarding, even the most advanced tools can go unused or misused.
“There’s often resistance to change and a lack of trust in AI,” Saputil said. “If employees aren’t involved or properly trained, AI becomes just another tool collecting dust.”
This problem is even more visible in the Philippines, according to Saputil, where enthusiasm for AI is strong but real adoption is still limited. While larger companies and government agencies are pushing forward with AI programs, many small businesses struggle to keep up.
This uneven progress has created what some experts call a “maturity gap.” It reflects the difference between the region’s potential for AI and how widely it is actually used. In the Philippines, this gap is tied to both infrastructure and education.
“To close the gap, we need more than just interest, we need consistent support,” Saputil said. “That means stronger digital infrastructure, better training programs, and real partnerships between the public and private sectors.”
AI strategy
Rather than choosing between proprietary and open-source tools, many companies are starting to mix both. Proprietary models offer reliability and are easier to use, but they can be expensive and may limit flexibility. Open-source models, on the other hand, offer customization and control, especially useful for working with local languages or sensitive data, but require more technical know-how.
A hybrid approach lets companies use ready-made AI tools for general tasks, while fine-tuning open-source models for more specific needs. For example, a company might use a proprietary model to run a customer service chatbot but rely on an open-source tool to analyze internal data in a secure way.
Successful AI adoption also depends on how well companies involve their workforce. Instead of using AI as a reason to lay off workers, businesses that include employees in the transition process often see better results. Upskilled teams help improve AI tools by offering feedback based on real-world use, making them more effective and easier to integrate.
“AI should support people, not replace them,” Saputil said. “The companies that work with their teams, invest in training, and keep learning as they go are the ones that get the most out of AI.”
Saputil noted that AI isn’t a one-time project; it requires ongoing care. Models can lose accuracy as data changes, and regulations may shift. Companies that build feedback loops and keep their teams updated are better equipped to adapt and improve their systems over time.
At the end of the day, success with open-source AI is not about saving money on software. It’s about building a strong foundation, technologically and culturally, that allows these tools to actually deliver value. For businesses in the Philippines, that means balancing cost, control, and the human side of innovation.
Get the latest before it trends. Follow Back End News on LinkedIn, Facebook, X, YouTube, and TikTok for updates and in-depth coverage across the tech and security landscape.