I work in the field of ai and cybersecurity at a small startup so can only speak from the experience of helping others solve some of these challenges. I think it comes down to setting guardrails (e.g you must have MFA to access X system), having the right monitoring in place (e.g automated scanners), and strongly leaning into ai-native tooling because ultimately machine-scale attacks and telemetry can only be countered with machine-scale defense.
Hire a CISO and let them build a team empowered to improve your security. Don't depend on a decentralized approach.
The most important aspect in my opinion is locking down permissions so that no one has access to everything and each team only has access to the components they need access to. Then one hack doesn't take down the whole company.
I work in the field of ai and cybersecurity at a small startup so can only speak from the experience of helping others solve some of these challenges. I think it comes down to setting guardrails (e.g you must have MFA to access X system), having the right monitoring in place (e.g automated scanners), and strongly leaning into ai-native tooling because ultimately machine-scale attacks and telemetry can only be countered with machine-scale defense.
Hire a CISO and let them build a team empowered to improve your security. Don't depend on a decentralized approach. The most important aspect in my opinion is locking down permissions so that no one has access to everything and each team only has access to the components they need access to. Then one hack doesn't take down the whole company.