Purpose
Clause 6 identifies every stakeholder and their attributes so the later steps can define the groups for which bias is examined.
What clause 6.1 asks
This step finds the people and groups touched by the system and the people and groups that touch it. Their attributes are recorded and then used to form the groups whose treatment is examined. The first set is produced here and revisited during the data, evaluation and risk work. This is a search, not a judgement: the evaluation comes later.
Why it matters
The identification and evaluation distinction matters. A team that evaluates while it identifies will drop the stakeholders it finds inconvenient or hard to reach. The clause asks for the full set first, with attributes, and leaves the judgement to the later steps. The requirement that the attributes are used to make groups is what turns a list of names into a testable population. Without it, a data map can claim representation without saying which group is represented.
What good looks like
Start with a wide list and record how each name was found. Include people who never interact with the system but are affected by its decisions, and people who shape the system without using it. For each, record the attributes that matter in this context. Do not score or rank them yet. Then hand the set to clause 7 with the attribute definitions, so the data map can be built against the same groups.
Evidence an assessor asks for
The assessor will ask how the team knows the set is complete. A register with no record of the search method invites the question of who was missed.
- Stakeholder register Every stakeholder and group found, with the route by which it was found
- Attribute list The inherent and context attributes for each stakeholder
- Group definitions The groups used for the bias examination, with the attributes that define them
- Identification log The sessions, sources and participants used to build the set
Where it commonly fails
- Only users and customers are listed
- Attributes are recorded for some stakeholders and left blank for the rest
- Groups are defined after the data is collected, so the data decides the groups
What to do before publication
Do the wide identification before any scoring. The evaluation belongs in clauses 8 and 9.
Read across
- 1.4 Field of application
- 5.2.2 Stakeholder identification in the life cycle
- 6.3 Outputs
- 6.4 Actions
- 7.4.2.3 Data map