Studies
Studies
You should be able to trust a result you did not run yourself. So each protocol is pre-registered before data exists, analyses follow the registered plan, null results are published, and an Advisory Board reviews the program from outside.
Labs
BTI’s work is organised into standing research groups and a flagship program. You can’t measure the future, but you can measure a difference.
Methodology
It is unsettling to defend against a capability no one has measured. Each study asks one question in one of two registers: how much AI changes what an attacker can do, and how much it changes what a defender can do.
All Studies
Agent vs. Scanner
Does AI-integrated reconnaissance find more of an environment's true exposures than a standard scanner at identical access?
Lure Quality at Capability Scale
As models get more capable, do the lures they produce get good enough that a low-resource attacker clears a professional bar?
Time to Insurable Posture
Does a defender reach a carrier-defined security baseline faster with AI assistance, verified control by control?
Alert Triage at Capability Scale
How much of an analyst's triage work can AI carry, and how fast is that share growing?
Compliance Gap-Closure
Does an AI-assisted team reach a real compliance standard faster than one working the established way?
