I’m more optimistic than ever about what AI can unlock for businesses and their customers. For businesses, AI drives efficiency, cost savings, strategic insight, and competitiveness. For customers, it delivers faster service, more personalized experiences, and often better value. But AI is also changing our risk profiles in subtle and sometimes not-so-subtle ways.
Todd Thorsen, our CISO, asked our experts to predict how AI would impact risk profiles in 2026, and they came up with 5 predictions (below), which we published in the blog, Recovery Confidence, Not Just Backups: AI Resilience in 2026: https://lnkd.in/eGwX5Z7K
1. Lower-noise attacks will matter more than high-drama events
2. Resilience is increasingly defined by the integrity of recoverable data
3. Agentic automation changes the shape of access and accountability
4. Training and tuning data becomes a high-value enterprise asset
5. Identity assurance will be pressured by better impersonation at scale
In this post, I’ll tackle the first prediction, and over the next few weeks, I’ll dive into each of the others.
Prediction 1: Lower-noise attacks will matter more than high drama events
Simply put, the threat profile is shifting from loud, obvious attacks to quiet ones that blend into normal operations. AI is helping attackers look legitimate, from using real credentials and familiar tools to highly believable social engineering. This makes attacks harder to spot.
In years past, there was most likely a single dramatic event. Now, it will be more subtle, with damage accumulating over time.
It’s not enough to just have backups. You need to know if you can trust them. If an attack goes unnoticed long enough, your backups might also contain corrupted data.
Your updated priorities should be validating backup integrity, enabling granular point-in-time recovery, strengthening audit trails to reconstruct what happened (and by whom), and conducting regular recovery tests so you can restore quickly and confidently.
This leads to number 2 - but that’s the next post.