Skip to learning paths
GLACIS Academy
Illustrative view of steel and glass buildings in a financial district at dawn

AI safety / runtime evidence / insurance

Glacis
Academy.

Understand the risk.
Ask for the right evidence.

A working education for the people building, assessing, and insuring AI.

Start with the market briefing
September 2026 editionChoose your path

The curriculum

Three ways in.
One connected picture.

Start where your next decision is.

01

The market conversation

Start with a loss scenario.

A concise briefing for underwriters, brokers, and the teams building AI controls.

Open the briefing
02

AI safety in practice

Follow the action. Test the control.

Agents, clinical reliance, drift, shared dependencies, and the evidence an incident leaves behind.

Explore the risk map
03

Insurance foundations

Build your insurance judgment.

Producer licensing, delegated authority, policy anatomy, and underwriting — with adaptive practice.

Continue learning

The question behind the evidence

What happened?
What can we establish?

Trace a workflow from its intended use to its observed effects. Keep the gaps visible.

Explore six workflows
  1. 01 / INTENTDefine the task & authority
  2. 02 / CONTROLTest the consequential boundary
  3. 03 / EVIDENCERecord the decision & result
  4. 04 / JUDGMENTEvaluate the record & its limits

Read. Practice. Revisit.

A reference you can
put to work.

91 lessons and 379 practice items, with source links and explicit trust markers. Work a case, test your understanding, then return to the evidence.

Work a case Read the reference Check the VERIFY radar

Lloyd’s Lab · Cohort 17

Glacis has been selected for Lloyd’s Lab. The market lab curriculum turns that opportunity into practical questions about loss, controls, evidence, and a bounded pilot.

Read the market lab