| Mandate |
Terms of reference for a threat assessment |
1 |
| Method |
Trope catalogue |
2 |
| Method |
Claim card |
3 |
| Method |
Constructive discussion format and crux map |
4 |
| Method |
Graded evidence record |
4 |
| Human mind |
Team working agreement for AI use |
6 |
| Human mind |
AI policy for a school or graduate programme |
7 |
| Human mind |
Sycophancy evaluation configuration |
8 |
| Human mind |
Safeguarding policy for an assistant |
9 |
| Society |
Content-provenance policy |
10 |
| Society |
Harm-path review of automated decisions |
11 |
| Society |
Disparate-impact test configuration |
12 |
| Society |
Oversight-quality metric |
14 |
| Society |
Dependency map with exit references |
15 |
| Society |
Model-routing and carbon policy |
17 |
| Misuse |
Callback-verification policy for approvals |
20 |
| Misuse |
Supplier-exposure record |
21 |
| Machines beyond control |
Shortcut-rate evaluation plan |
22 |
| Machines beyond control |
Layered evaluation plan with monitoring and decoys |
23 |
| Machines beyond control |
Stop-path test record |
24 |
| Machines beyond control |
Agent incident timeline record |
25 |
| Machines beyond control |
Delegation register with reversibility test |
26 |
| Response |
Supplier evidence request from a safety framework |
29 |
| Response |
Control-to-threat coverage matrix |
30 |
| Individuals |
Household and classroom agreement on AI use |
31 |
| Response |
AI threat register and schema |
32 |
| Response |
One-page board briefing |
32 |