Cognitive load
Does the system add information, decisions, prompts or alerts faster than people can reliably process them?
42005.ai connects AI governance with human factors, work design, occupational hygiene, safety and the realities of how people actually perform work.
Independent professional initiative · Not affiliated with or endorsed by ISO or IEC
A system can be technically correct and still create unsafe work.
Automated scheduling, computer vision, algorithmic management and generative AI can change pace, attention, autonomy, workload and decision-making.
That means AI implementation is not only a technology question. It is also a work design and human performance question.
Assess what changes when an AI system enters real work: who is affected, what demands shift, what can fail, and whether people retain meaningful control.
Does the system add information, decisions, prompts or alerts faster than people can reliably process them?
Will workers trust the recommendation because it came from the system?
Optimisation can increase pace, pressure and fatigue even when productivity improves.
More alerts do not necessarily mean better awareness.
People need time, information, competence and genuine authority to challenge, pause or override a system.
Repeated reliance can change skill retention, situational awareness and professional judgement.
AI changes the relationship between people, information, tasks, teams and technology. A meaningful impact assessment therefore needs a human-centred design lens from the start.
Mental workload, attention, decision-making, interruption, fatigue and information presentation.
Trust, reliance, explainability, feedback, error recovery and the quality of human oversight.
How AI changes job content, pace, autonomy, staffing, coordination, recovery and performance expectations.
Workers are not merely end users. Their experience is evidence for identifying foreseeable impacts and designing controls.
Digital decisions can change exposure to physical, ergonomic, chemical and environmental hazards in the real workplace.
A deeper guide to applying ergonomics and human factors thinking to AI-enabled work.
Open the HFE guide ↗42005.ai explores how AI system impact assessment can be translated into practical workplace questions for safety, health and human factors professionals.
View the ISO standard information ↗A workplace impact assessment should follow the change from system to task to people to consequences — then return to controls and monitoring.
Purpose, capability, limits and foreseeable use.
Tasks, pace, staffing, workflow and environment.
Attention, skill, trust, wellbeing and authority.
Benefits, harms, uncertainty and affected groups.
Design, verification, escalation and safeguards.
Feedback, near misses, overrides and change triggers.
Register AI systems, document workplace impacts, record evidence, assign controls, compare worker experience with management assumptions, and generate a decision-ready assessment report.
Browser-based V4 · Local persistence · Export / import workspace data
18 questions. About five minutes. Designed to surface workplace impact signals that deserve deeper review.
Describe the system in the context of the work it changes.
Zephan Chan is a workplace health, ergonomics and occupational hygiene practitioner whose work increasingly focuses on the human and operational impact of AI. Through 42005.ai, he is developing practical approaches that connect AI governance with human factors, worker wellbeing and real work systems.
The aim is to help lead a more practical conversation: not only “Does the AI work?”, but “What does the AI do to work, people and risk?”
Zephan is bringing together professionals from workplace safety, occupational hygiene, human factors, design, technology and AI governance to explore how AI impact assessment should work in real workplaces.