Mozilla’s CTO thinks AI should be built like the internet

What changed

Mozilla’s CTO Raffi Krikorian outlined a vision for AI development modeled on how the internet was built: open, interoperable, and customizable. He noted that while consumer AI tools like ChatGPT and Claude attract massive user bases, many businesses are shifting toward open-source AI models. These models give organizations control to tailor AI to their specific needs, rather than relying on closed systems controlled by a few providers.

Why builders should care

Open AI models let builders avoid lock-in and keep sensitive data in-house, which is crucial for industries with privacy or compliance demands. Companies can customize training datasets, tweak model behavior, and integrate AI more deeply into existing infrastructure. This lowers dependence on big cloud providers’ closed platforms and offers more transparency about how AI systems operate.

The practical takeaway

For AI operators and founders, investing in open models can reduce licensing costs and create competitive advantages through unique AI capabilities. These architectures shift power away from a handful of AI giants toward a more distributed ecosystem, increasing innovation but also raising the technical bar for implementation and ongoing maintenance. Decision-makers should weigh these trade-offs carefully.

What to watch next

Monitoring how enterprises balance open-source AI adoption with proprietary models will reveal which approach delivers better outcomes at scale. The availability of tools to manage open AI model deployment securely and efficiently will directly impact how fast this trend grows. Expect vendors to emerge that specialize in bridging open AI infrastructure with enterprise requirements, simplifying integration and compliance.

AI Quick Briefs Editorial Desk