Key words

Word 1

SANDBOX

• Supervision to ensure rules and safety are followed.
• An official order demanding documents or testimony.
• A safe test environment separated from real systems.
• Able to act and make decisions without direct human control.

Word 2

ESCALATE

• Supervision to ensure rules and safety are followed.
• To raise an issue to higher levels for action.
• A safe test environment separated from real systems.
• To sort and prioritise issues for response by urgency.

Word 3

AUTONOMOUS

• Supervision to ensure rules and safety are followed.
• An official order demanding documents or testimony.
• Able to act and make decisions without direct human control.
• A safe test environment separated from real systems.

Word 4

MISALIGNMENT

• Behaviour by an AI that conflicts with intended goals.
• Able to act and make decisions without direct human control.
• A safe test environment separated from real systems.
• An official order demanding documents or testimony.

Word 5

OVERSIGHT

• An official order demanding documents or testimony.
• A central place where software or data is stored and managed.
• To raise an issue to higher levels for action.
• Supervision to ensure rules and safety are followed.

Word 6

TRIAGE

• To raise an issue to higher levels for action.
• Able to act and make decisions without direct human control.
• To sort and prioritise issues for response by urgency.
• Behaviour by an AI that conflicts with intended goals.

Word 7

REPOSITORY

• A central place where software or data is stored and managed.
• Able to act and make decisions without direct human control.
• An official order demanding documents or testimony.
• To sort and prioritise issues for response by urgency.

Word 8

SUBPOENA

• An official order demanding documents or testimony.
• Behaviour by an AI that conflicts with intended goals.
• A central place where software or data is stored and managed.
• A safe test environment separated from real systems.