Microsoft Addresses Public's Biggest Fear About Using AI

Microsoft has emphasised that it does not use customer data to train the artificial intelligence (AI) models it provides. The affirmation was delivered by Microsoft ASEAN Senior Cloud and AI Platform Go to Market Lead, Fiki Setiyono, during the Work Trend Index 2026 Indonesia event in Jakarta on Tuesday (30/6/2026).

Fiki stated that Microsoft can technically ensure customer data is not sent to retrain the AI models. “This is what I might need to help elaborate on, data is not sent to retrain the relevant reasoning model,” Fiki said.

He cited the Azure OpenAI service, one of Microsoft’s frontier models, as an example. According to him, data processed through the service is not used to enhance the AI model’s capabilities. “This means that customer data using our technology is not used to make the reasoning model smarter. Technically, we establish that in a way that can be proven technically,” he explained.

On the other hand, Fiki noted that industry demand for AI is now evolving towards sovereign AI, a concept that emphasises data governance and AI infrastructure sovereignty within each country or organisation. He stated that the growing discussion around sovereign AI is not a setback for the industry, but rather proof that AI adoption is maturing.

“We also see, and this is actually encouraging, that the industry’s need for AI is now moving towards sovereign AI,” Fiki said. He added that for Microsoft, sovereign AI is not just a concept. The company has introduced support based on two main principles.

The first principle is ensuring the availability of AI infrastructure. He mentioned that Microsoft has established a data centre in Indonesia equipped with AI infrastructure, including VMs and GPUs that customers can utilise. “The existence of Microsoft’s data centre in Indonesia already has AI-based infrastructure, VMs, GPUs that can be utilised. That is the first principle of how Microsoft wants to democratise the existence of infrastructure to support sovereignty,” he stated.

The second principle is expanding access to various AI models. According to Fiki, Microsoft does not limit users to just one AI model, allowing customers to choose the model that best fits their needs. “Democratisation is not only in infrastructure but also in AI models. At Microsoft, we do not limit AI users to just one AI model,” he concluded.