Drafting and structure
AI tools may support drafting, summarising, editing and organising information.
Human judgement remains responsible for the meaning, tone, usefulness and accuracy of the work.
AI Use & Transparency
AI is useful when it is kept in context.
FabUX uses AI tools carefully, practically and with human review.
AI can help with drafting, analysis, coding and structure. It can also fill gaps too confidently when the signals are weak.
The work still depends on observation, interpretation, trust and judgement.
How AI may be used
Support, not substitution.
Analysis support
AI can help sort material, identify patterns and test alternative ways of explaining something.
FabUX does not treat generated output as observed behaviour. Observation, context and human interpretation still matter.
Coding and prototyping
AI tools may assist with code, prototypes, documentation and technical checks.
Outputs are reviewed before use. Helpful acceleration is not the same as automatic correctness.
Human judgement
FabUX values observation over assumption.
Behavioural interpretation, contextual understanding, trust and ambiguity all need human judgement. AI can support that work, but it should not replace it.
Clarity matters
AI can expose ambiguity as quickly as it resolves it.
When an organisation is unclear, AI systems may try to complete the missing context. That is why answer engine readiness is also a clarity, content and trust problem.
Assumptions and gaps
AI systems naturally bridge gaps with assumptions.
That can be useful when exploring ideas. It can also create confident answers from weak, incomplete or ambiguous information.
Hallucinations
AI can accelerate understanding. It can also accelerate misunderstanding.
Hallucinations often point back to knowledge gaps, fragmented content, inconsistent messaging or unclear explanations of what an organisation is and is not.
Privacy-aware use
FabUX avoids putting sensitive, unnecessary or participant-identifiable information into AI tools without a clear reason and appropriate handling.
The same principle applies as wider research work: collect less, explain more and respect the context.
Review and oversight
AI-assisted work is reviewed by a human before it is treated as useful.
The final responsibility sits with people, not with a tool that cannot understand the full situation by itself.