My takeaways from the 2025 Microsoft Swiss Security Day: Are We Ready for a Sustainable AI Future?

The event showcased Microsoft’s advancements in AI for security, including their latest tools and plans for IT professionals. I went in without any specific expectations, but I left with ideas I can bring into my daily work.

One topic in particular gave me a lot to think about—not as a criticism of Microsoft or their tools, but as a personal observation and concern.

They presented a scenario where junior and senior SOC analysts use Copilot to summarise and triage issues within a cybersecurity tool. The idea is that Copilot can help juniors become productive more quickly by leveraging knowledge that would typically take months to acquire. This could benefit organisations by enabling faster onboarding and improved productivity.

However, I see a potential downside: it may slow down these juniors learning journey. If they rely too heavily on AI to interpret and summarise information, they might miss out on developing a deep understanding of the subject.

For senior SOC analysts, who already have a strong grasp of the domain, these summaries might not even be necessary. But if we follow this path, I worry that entry-level roles may become increasingly rare as they are automated away. And if we lose those roles, how will we cultivate the next generation of experienced professionals?

This concern ties into the broader topic of agentic AI systems—technologies that execute tasks under human supervision and are becoming a major focus now and in the near future. But who will those supervisors be if we don’t invest in developing human expertise? Without that, we risk having fewer qualified professionals to oversee these systems.

While AI tools can certainly boost productivity, they also make processes more abstract. To truly understand and manage them, we need deep knowledge of the field. From my experience working with trainees who use Copilot to learn software development, I’ve seen both sides of the coin. On one hand, they can build things quickly and appear productive. On the other, the burden often falls back on senior staff to correct mistakes and teach fundamentals that weren’t learned because the trainees skipped the routine and repetition needed to truly master the craft.

This isn’t just a technical issue; it’s an organisational and ethical question. How do we want to use these tools today and in the future? What kind of learning culture do we want to cultivate? Even as AI improves, I believe human supervision will become more important, not less. I’ve also noticed that junior or entry-level jobs in the software industry are becoming harder to find, with even higher requirements for applicants and their tasks.

But I also had a great conversation with someone who brought one of their trainees along. She was a young woman working in a large, well-known Swiss organisation. Hearing her speak with such passion and pride about her role and the support she receives gave me hope. It reminded me that with the right mentorship and learning environment, the next generation can grow—even in a world shaped by AI.

Here’s the real question we need to ask ourselves:

Are we preparing for a future that’s not just efficient, but sustainable? One where productivity is aligned with mentorship, and automation goes hand in hand with human growth?

It’s a conversation worth having—sooner rather than later.