Data, Rules and Human Decisions
A decision can be reviewed only when its evidence, authority and consequences remain visible.
Before a university assigns a scarce resource, it needs to know where the information came from, what it can rely on and what still needs checking.
Rules set the boundaries of permissible action. Analysts, subject specialists and proposed AI assistants can prepare options. The decision belongs to a person with the authority to make it.
This article develops an author's research model, with an illustrative laboratory conflict. Its AI functions and institutional decision procedures are proposed concepts, not evidence of a live University 5.0 service.
- Responsible role
- Data owner
- Produced evidence or action
- State record
- Condition for moving on
- Source and owner are known
Showing 5 relevant stages out of 11. These describe a proposed workflow, not recorded approvals.
A decision passes through several distinct kinds of review
Data describes the state of a situation. Analysis identifies relationships. Rules describe permissible action. People with defined responsibilities examine alternatives and their consequences.
The operating sequence contains twelve steps; seven architectural levels group them by purpose. A separate worked example follows the whole sequence.
- 01Source
- 02Data
- 03Verification
- 04Observation
- 05Interpretation
- 06Scenarios
- 07Rules
- 08Agreement
- 09Decision
- 10Action
- 11Outcome
- 12Revision
Examine decision governance in University 5.0
A separate interactive model explores the same ideas. Its interface is currently in Russian.
Open the model (Russian)Seven functions support an accountable decision
The sequence matters, but its functions have different responsibilities: evidence, quality, interpretation, rules, options, a human choice and revision after the outcome.
Sources and data
Responsibility: Data owner
- Source
- Timestamp
- Collection method
- Version
- Owner
- Scope of applicability
Expected output: Traceable dataset
An important conclusion must be connected to its source
Useful evidence records its origin, collection method, owner, version and scope. A statement without these cannot be assumed to apply to every decision.
Digital twin data
- Origin: Campus and equipment model
- Collection: Automated collection of state information
- Owner: Digital-twin service
- Scope: Operational view of resources
Learning-task outcomes
- Origin: Teaching modules and practical activities
- Collection: Recording solutions and feedback
- Owner: Learning programme
- Scope: Educational context of the decision
Project data
- Origin: Project teams and partners
- Collection: Project reports and schedules
- Owner: Project coordinator
- Scope: External commitments
Expert opinions
- Origin: Subject specialists
- Collection: Reasoned expert assessment
- Owner: Expert
- Scope: Applicability and consequences
Rules and institutional documents
- Origin: Institutional procedures
- Collection: Approval by the authorised governance role
- Owner: Governance role
- Scope: Limits of permissible actions
External events and sources
- Origin: Partners, publications and wider environment
- Collection: Recording in an event log
- Owner: Event coordinator
- Scope: Context of change
Trace a proposed conclusion back to its grounds
Illustrative proposal: allocate the laboratory to the external project on Thursday. The example contains possible evidence records; it does not supply the signed documents required to verify this proposal.
External commitment
- Responsible owner
- Project coordinator
- Required evidence
- Signed schedule
- When it applies
- Current week
- Limitation
- Fixed deadline
- Effect on the proposal
- Sets the priority of a time window
Data, facts, observations and recommendations are different
- 01DataA laboratory is shown as available for four hours.
- 02FactThree processes request the same resource.
- 03IndicatorCombined requested time exceeds the available window.
- 04ObservationThe resource conflict affects three university activities.
- 05InterpretationRescheduling one activity can delay results for the others.
- 06ScenarioSeveral options for allocating the laboratory are possible.
- 07RecommendationPriority may depend on critical-path effects, equipment uniqueness and the cost of rescheduling.
Statements shown as confirmed are assumptions within the illustration. An actual decision requires access to independently reviewable source evidence.
How data can lead to a recommendation
Data
A laboratory is shown as available for four hours.
- Transformation
- Recording a state
- Assumption
- Digital-twin readings are accurate
- Uncertainty
- Assumed verified for this example only
- Reviewing role
- Data owner
- Result
- State record
A responsible decision records what is not yet known
Sources of uncertainty
- Incomplete data
- Information becoming outdated
- Conflicting sources
- Divergent expert assessments
- Several possible causes
- Unknown consequences
- Changing context
Qualitative states
- Data verified
- Data needs clarification
- Multiple explanations possible
- Decision depends on conditions
- Expert review needed
- Reassessment needed
What is known, and what needs a decision
A research experiment, an external project and a learning module request the same laboratory. The following are categories within the hypothetical example, not observed live resource data.
Known in this example
- Available time
- Tasks
- Current availability
- Deadlines
- Equipment
Rules describe what may be done
A useful rule states its trigger, the permitted or required action, its limits and who can authorise an exception. The following are illustrative rules for the research architecture, not confirmed institutional policy.
Access rule
- Condition
- Request for data or a resource
- Action
- Check permissions
- Limit
- Scope of the requesting role
- Accountable role
- Data owner
- Exception conditions
- Emergency access requires a separately authorised procedure
Priority rule
- Condition
- Conflicting requests
- Action
- Compare critical-path consequences and uniqueness
- Limit
- Does not cancel existing commitments
- Accountable role
- Process lead
- Exception conditions
- Escalate for governance approval
Coordination rule
- Condition
- A decision affects several processes
- Action
- Collect necessary approvals
- Limit
- Full approval route
- Accountable role
- Coordinator
- Exception conditions
- Expedited route only where explicitly authorised
Safety rule
- Condition
- Risk to participants or data
- Action
- Stop the unsafe action
- Limit
- Takes precedence over competing priorities
- Accountable role
- Governance role
- Exception conditions
- No exception for bypassing this safety condition
Data-use rule
- Condition
- Handling participant data
- Action
- Grant minimum necessary access
- Limit
- Within the authorised task
- Accountable role
- Data owner
- Exception conditions
- Any audit access also needs its own authority
Exception rule
- Condition
- Unanticipated case
- Action
- Record the proposed exception
- Limit
- Requires approval
- Accountable role
- Governance role
- Exception conditions
- Record reasons and decision in the log
Reassessment rule
- Condition
- A decision has consequences
- Action
- Set a review checkpoint
- Limit
- Defined review date or condition
- Accountable role
- Coordinator
- Exception conditions
- Any extension must be justified
How a rule changes the options
Resource uniqueness
- Possible response
- Consider priority for a process without an equivalent substitute
- Affected work
- Teaching module, investigation and external project
- Limits
- Assess whether an alternative is genuinely equivalent
- Required review
- Resource specialist
- Reassessment
- After the first use
Who can act, and who receives the result
Data owner
- Accessible information
- Relevant datasets
- Permitted actions
- Provide · clarify · correct · confirm source
- Required coordination
- Analyst
- Output
- Traceable dataset
- Limit
- Cannot choose the operational response
- Next role
- Analyst
Different assumptions lead to different possible actions
Four ways of allocating the laboratory are examined. Each affects research, teaching, external obligations and risk in a different way. All remain proposals until reviewed and authorised.
Scenario A. Allocate the laboratory to research
- Grounds
- Equipment uniquely needed for the experiment
- Rules
- Resource uniqueness · critical path
- Potential benefit
- Research work can continue
- Trade-off
- External project must be rescheduled
- Risks
- Commitment and reputational risk with the partner
- Required roles
- Researcher and programme leader
- Review point
- After the first experimental iteration
Scenario B. Allocate the laboratory to the external project
- Grounds
- Confirmed external commitment
- Rules
- External commitment · critical path
- Potential benefit
- Partner commitment can be met
- Trade-off
- Research and teaching move
- Risks
- Scientific results may be delayed
- Required roles
- Project coordinator and programme leader
- Review point
- At the end of the external project window
Scenario C. Divide the laboratory time window
- Grounds
- Some tasks may be compatible
- Rules
- Minimise rescheduling · coordination
- Potential benefit
- Each of the three processes may keep some time
- Trade-off
- Coordination becomes more complex
- Risks
- At least one activity may lose quality
- Required roles
- Coordinator and laboratory head
- Review point
- Halfway through the shared window
Scenario D. Use a digital alternative
- Grounds
- A suitable virtual resource may be available
- Rules
- Minimise rescheduling · educational checkpoint
- Potential benefit
- Educational work can continue
- Trade-off
- Less hands-on experience
- Risks
- The digital substitute may not be sufficient
- Required roles
- Programme leader and resource specialist
- Review point
- At the next in-person stage
Compare two distinct scenarios
Choose one alternative in each column. Every criterion is qualitative; no score or winner is invented.
First scenario
Second scenario
- Critical path
Scenario A: Favours the research critical-path candidate
Scenario B: Favours the external project if its dependencies are confirmed
- Resource uniqueness
Scenario A: Strongly favours scarce experimental equipment
Scenario B: Requires resource-expert assessment
- Deadlines
Scenario A: Delays the external commitment
Scenario B: Preserves the partner deadline
- Educational outcome
Scenario A: Teaching-module rescheduling needs agreement
Scenario B: Needs agreement on rescheduling
- Research outcome
Scenario A: Favours continuation of research
Scenario B: Risks delaying the experiment
- External commitments
Scenario A: Creates additional contractual pressure
Scenario B: Strongly favours the confirmed commitment
- Cost of rescheduling
Scenario A: Moves the external project
Scenario B: Moves research and teaching
- Reversibility
Scenario A: Keeps several options after the research iteration
Scenario B: Depends on a later reassessment
Specialists examine a proposal before approval
The expert's conclusion explains the application of rules, limits and likely consequences. It does not replace a separate decision by the responsible person.
- Do the rules apply to these circumstances?
- Is the chosen option feasible?
- What consequences might affect linked activities?
- Which assumptions must remain true?
- When must the result be checked again?
A responsible person weighs the evidence and chooses the action
The decision must be grounded in evidence that can be reviewed, rules that apply to the situation, expert conclusions and likely effects on connected work.
- Situation and traceable evidence
- Explicit uncertainty
- Applicable rules
- Alternatives considered
- Expert conclusions
- Proposed or authorised scenario and accountable role
- Conditions and review checkpoint
Preparing or selecting a proposed decision on this page grants no permission and does not alter a schedule or any institutional record.
Work through the decision without confusing a proposal with approval
Choose one of the four scenarios. Each step reveals which question still needs attention; the final proposal remains illustrative until the appropriate people confirm the evidence, roles and permissions.
Examine the evidence
Identify each request, its source, timetable, resources and external commitments. Missing evidence must remain marked as missing.
- Scenario under consideration
- B · Allocate the laboratory to the external project
- Grounds requiring verification
- Confirmed external commitment
- Relevant proposed rules
- External commitment · critical path
- Potential benefit
- Partner commitment can be met
- Trade-off
- Research and teaching move
- Risk to review
- Scientific results may be delayed
- People to involve
- Project coordinator and programme leader
- Review checkpoint
- At the end of the external project window
The selection is a locally displayed proposal only. No real decision, approval, booking or resource change has been made.
Actual consequences must be checked against the decision
Implementation generates new evidence only after an authorised action occurs. The model calls for recording both intended outcomes and unexpected changes.
- Actual use of the laboratory
- Meeting stated deadlines
- Effects of rescheduling linked activities
- Quality of the digital substitute
- Research outcome
- Educational outcome
- New constraints and unresolved questions
Consequences can justify changes, subject to approval
An outcome may justify a proposed revision to these records or rules. Neither this article nor an analytical recommendation automatically changes them.
- Priority rule
- Description of the resource
- Digital twin
- Rescheduling-cost model
- Timetable
- Analytical observations
- Next decision cycle
One situation looks different to different responsible people
Analyst
Question: What does the evidence show and which interpretations remain plausible?
Relevant information:
- Sources
- Quality
- Dependencies
- Contradictions
- Scenarios
The proposed decision cycle connects evidence, tasks and accountability
The University 5.0 model describes several complementary systems. The links below lead to the corresponding research material; they are not evidence that the full architecture is operational.
EduOS Lad
Would coordinate:
- Tasks
- Roles
- Rules
- Approvals
- Review checkpoints
Digital twin
Would record:
- States
- Resources
- Relationships
- Events
- Decisions
- Consequences
Educational analytics
Would provide:
- Observations
- Interpretations
- Scenarios
- Uncertainty assessments
Multi-agent environment
Would assist with:
- Analysis
- Verification
- Critical review
- Coordination
- Preparing alternatives
Research reproduction cycle
Would use reviewed results for:
- Revising research
- Changing educational modules
- Developing projects
- Producing new evidence
Sixteen stages from conflicting requests to the next decision
Three activities need the same laboratory. Filter the example by evidence, rules, roles, decisions or outcomes. Every stage is illustrative and may depend on evidence or approval not supplied here.
Resource conflict
Three processes request one laboratory in overlapping time windows.
Ten principles of accountable decisions
Traceable provenance
Every important conclusion can be followed to a source and a record of changes.
Different states of meaning
Data, observations, interpretations, recommendations and decisions retain different statuses.
Explicit rules
Permitted actions, constraints and authority boundaries are stated in advance.
Role-based access
Participants receive only the data and functions authorised for their tasks.
Visible uncertainty
Missing evidence, contradictions and alternative explanations remain visible.
Distinct options
Materially different courses of action are considered before a decision.
Expert review
Professional applicability and consequences receive a separate assessment.
Human accountability
An authorised person selects an action and remains responsible for the consequences.
Decision record
Reasons, rules, approvals, responsible people and outcomes can be traced.
Learning from consequences
Actual outcomes inform data corrections and proposed rule or process changes.
What makes a decision accountable
- 01DataDecisions begin with information whose origins and quality can be examined.
- 02InterpretationAnalysts describe observations and scenarios using explicit methods and assumptions.
- 03UncertaintyMissing information and alternative explanations remain part of the decision record.
- 04RulesPermitted actions, restrictions and approval procedures are stated explicitly.
- 05AuthorityEach participant acts within a defined scope of responsibility.
- 06OptionsSeveral explainable scenarios are prepared with consequences and limits.
- 07DecisionAn authorised person approves an action and accepts responsibility for its consequences.
- 08RevisionActual results inform the next assessment and any authorised changes to rules or system data.
The proposed cycle would let participants trace a decision from source evidence and rules to a person who has authority, an action and its observed consequences. Actual system use would still depend on implementation, operational evidence and accountable institutional procedures.