I think the AI vs AI cybersecurity debate is missing the real risk

Everyone’s talking about attackers using AI vs defenders using AI.

But after working closely with teams at Periscope Technologies Inc, I’m starting to feel like that’s not where things are breaking.

What we’re seeing more often:

AI inside companies already has:

• Access to sensitive data
• Ability to trigger workflows
• Decision-making power

But very little verification around:

👉 What it’s doing
👉 Why it’s doing it
👉 Whether it should be doing it

Traditional security made sense when:

• Users were human
• Behavior was predictable
• Access was controlled

AI doesn’t fit that model at all.

Feels like we’re focusing heavily on external AI threats…

While ignoring a new category of risk:

Unverified AI execution inside systems

Curious if others here are seeing this?

Or is the focus still mostly on external threats?

periscope-tech.com