Code Blue / SiteGuard

Academic Industry Challenge · Arup

Can AI support worker safety without becoming another form of surveillance?

ROLE AND AUTHORSHIP

Five-person academic team. Personal contributions: senior HR manager interview, secondary research, synthesis, report sections, presentation slides 18–21, and shared visual direction.

Proposal

Research

Service strategy

Prototype

Feedback & limits

Original presentation photograph — speculative service-design proposal.

PROPOSED INTERVENTION / SPECULATIVE ACADEMIC CONCEPT

A safety system proposed around the workday.

The project investigated safety and wellbeing concerns among migrant construction workers in the UAE. The proposed SiteGuard service combines a fixed safety pole, environmental sensing and a supervisor dashboard.

01 / RESEARCH

Care, control and the workday.

An interview with a senior HR manager, secondary research and collaborative synthesis informed the proposed service. My contributions included the interview, research synthesis and documented sections of the final report and presentation.

24-hour worker-wellbeing journey — identified presentation contribution.

Stakeholder context — collaborative academic research synthesis.

From research to proposed decisions.

Heat, dust and fatigue

→ Proposed environmental hazard monitoring and safety alerts.

Communication barriers

→ Simplified visual warnings; multilingual support remains future development.

Supervisor decision complexity

→ Clear hazard identification, information hierarchy and response actions.

Surveillance anxiety

→ Transparent data use, consent, privacy safeguards and non-punitive governance are necessary.

These relationships describe proposed design decisions, not proven safety benefits.

02 / STRATEGIC FORESIGHT

A proposed system of interventions.

The speculative fixed safety pole, environmental sensing and supervisor dashboard were explored through a workday intervention model, three horizons and a proposed service blueprint.

Research raised questions around heat exposure, workplace hazards, reporting barriers, and the tradeoff between care and control. A three-horizon strategy explored possible development from monitoring toward wider wellbeing support.

Proposed fixed safety pole → environmental and hazard monitoring → on-site visual/audio warnings → supervisor dashboard assessment → response or emergency escalation. These are service relationships, not validated hardware capabilities or working emergency integrations.

Three-horizon strategy — identified presentation contribution.

Proposed service blueprint; no operational deployment.

03 / SIMULATED EXPERIENCE

Making the service visible.

The interfaces simulate a safety dashboard and hazard alerts. They do not demonstrate working AI or deployed sensors. Worker details are obscured in the supplied public-oriented previews.

Hazard → assessment → response. Each simulated screen identifies a risk, shows its context and offers supervisor actions. The three scenarios use the original interface artwork; worker identifiers remain obscured.

Simulated heat-stress alert; identifying details obscured.

Slip Hazard Alert — simulated prototype; worker identifiers obscured.

Simulated fall-risk alert; identifying details obscured.

ORIGINAL TEAM PRESENTATION / 29 SLIDES

The complete project presentation.

Explore the original final presentation: research, service strategy, prototype development and proxy feedback.

1 / 29 Code Blue · Beyond Efficiency

SCOPE & EVIDENCE

The team produced a service proposal, foresight roadmap and simulated dashboard screens. Early feedback came from peers, team members and an instructor acting as a client proxy.

No working AI deployment or on-site validation. Worker privacy, consent and identification require further resolution. Arup was an industry challenger in the course brief.

Peers, internal team members and an instructor acting as a senior client/supervisor proxy reviewed the mockups. Feedback highlighted alert clarity and recognizable icons, while asking for more nuanced supervisor actions, multilingual access, worker privacy and transparent data use. No construction-worker field trial, deployed AI performance or measured incident reduction was established.

Early proxy feedback — classmates, team members and instructor/client proxy.

DES475 academic challenge referencing Arup; no commercial commission, endorsement or implementation is claimed.