Not the same as
Value alignment
A component can respond predictably to intervention without endorsing the objective behind that intervention.
An adjacent property
Alignability
Can the system be reliably redirected when circumstances or higher-level goals change?
Alignability is the capacity for behavior to be redirected through interventions in the system's environment or architecture without requiring its underlying objectives to become the same as ours.
Try seeing it
Sometimes you do not change what the component wants. You change what its world rewards or permits.
A profit-seeking firm can change what it produces when relative prices change. It can alter behavior in response to taxes, contracts, regulation, ownership, or resource constraints while remaining profit-seeking.
The important property is therefore relational: does the surrounding architecture provide dependable handles through which competent behavior can be redirected?
Not the same as
Current good behavior
A system may behave well today while being extremely difficult to redirect when conditions change.
Look for
Live control handles
Which prices, permissions, dependencies, signals, rules, or intervention channels remain capable of changing behavior?
A question to carry elsewhere