AI is only meaningful if it delivers practical, tangible benefits to users. One of our customers, a large public health institution, needed live captioning and real-time translation to reduce barriers between caregivers and patients, to make video meetings as inclusive as possible. AI in video conferencing can solve this…but this isn’t the kind of generative AI that’s making all the headlines these days. This is a kind of AI that takes user actions and transposes them into something meaningful for others. It’s AI that allows us to solve an important problem for our users regarding the accessibility of their critical communications.
Once we determined with our customer that ‘yes, AI can solve this for you’, then we were hit with the next question: how do we make all this newly generated AI data sovereign or private?
The whispers around AI sovereignty have grown a tiny bit louder lately, making their way from the quiet corners of government offices and into corporate corridors across Europe especially. A recent article in CNBC explains that the increased focus on sovereign AI solutions is due to the growing number of European organizations seeking ways to reduce their dependency on non-European providers, ensuring that their AI data stays confined to walls of their organization and/or a sovereign cloud service within the borders of their country or region.
While this trend is undeniably important and one that I expect will grow globally with time, I am still classifying this as a ‘loud whisper’ from the market – as the focus on AI sovereignty still resides in niche pockets outside the mainstream. For our customers, AI sovereignty is not necessarily the first question they ask when inquiring about our secure video conferencing solutions. The first step is always about fixing the problem at hand – the problem that AI can potentially solve for them. And then we dive into how we can deliver that in the best way possible, to meet their specific needs.
For industries working with sensitive health data, such as this customer, or legal issues, or perhaps even information that may have consequences for national security, there is simply no option to store this AI data in a cloud service residing an ocean away. Sometimes, it’s not even an option to store it in a cloud service at all – it may not even be able to leave the organization’s walls. While this problem is secondary to the initial problem, it becomes just as important – as now we are dealing with the challenge of ensuring that the AI data is private and stays protected.
AI data sovereignty and privacy are and will become increasingly expected from our customers, especially in regions like Europe and the US. They will want a solution that solves their problems, but they will expect it to be delivered in a way that aligns with their regulatory, privacy, and operational requirements. Whether that means self-hosted deployments, leveraging sovereign cloud solutions, or exploring alternatives that ensure compliance with strict data protection laws – organizations need flexibility in how they adopt AI.
For AI providers, including companies like us delivering AI in video conferencing, the lesson is clear and becoming clearer by the day: sovereignty matters. At Pexip, we’re committed to providing solutions that solve real problems for our customers. And at the same time, we recognize that delivering these solutions in a sovereign, flexible, and compliant manner is no longer “nice to have”; it’s a fundamental part of meeting the expectations of a rapidly evolving AI landscape. For us, solving our customers challenges and delivering sovereign solutions aren’t competing priorities – they’re two sides of the same coin.