By Mohamad Nur Hafidz Bin Ahmad Fuaad
Introduction
Artificial Intelligence (AI) has evolved from a futuristic concept into a practical tool shaping multiple industries, including healthcare, manufacturing, finance, and now — safety management. With its origins dating back to the 1950s, when Alan Turing first posed the question, “Can machines think?”, AI has undergone decades of transformation. From rule-based systems to today's advanced machine learning and computer vision technologies, AI has reached a point where it can recognize patterns, learn from data, and make autonomous decisions with remarkable accuracy.
In the context of laboratory safety, particularly in high-risk environments like civil engineering construction labs, AI offers a powerful toolset. Traditionally, safety has relied heavily on human observation, physical signage, and routine checklists. While these methods remain important, they are often reactive — addressing hazards after they occur. AI introduces the possibility of a proactive, predictive, and preventative approach to safety.
The idea of integrating AI with lab safety stemmed from the growing need to enhance hazard recognition capabilities, reduce human error, and foster a culture of continuous monitoring. With high-definition surveillance cameras, real-time analytics, and predictive algorithms, AI is now capable of identifying risks in real time and alerting personnel before accidents happen.
Benefits of AI in Lab Safety
Incorporating AI into lab safety protocols delivers a wide range of benefits. Below are three major advantages:
1. Proactive Hazard Detection
AI can scan environments using surveillance cameras and IoT sensors to detect potential hazards before they become dangerous. Whether it’s a chemical spill, an exposed wire, or a loose scaffold, the AI can identify the anomaly and issue a warning instantly. This proactive capability is key to preventing minor issues from escalating into major incidents.
2. Real-Time Response and Monitoring
Unlike traditional checklists that depend on human attention and scheduling, AI offers 24/7 monitoring. In high-risk lab environments, real-time alerts can mean the difference between safety and injury. AI-powered platforms can notify safety officers or pause experiments remotely if thresholds are exceeded or risks are detected.
3. Data-Driven Safety Culture
AI collects vast amounts of data on near misses, user behavior, and environmental conditions. These insights help institutions refine their safety protocols, identify training gaps, and develop targeted interventions. This promotes a data-driven culture where decisions are backed by analytics rather than guesswork.
Real Life Examples
Case Study 1: HDB Construction Sites — AI Monitored CCTV Hazard Detection
Overview: In 2019–2024, Singapore’s Housing & Development Board (HDB) mandated AI-powered CCTV monitoring at numerous public housing construction sites—such as at Tanglin Halt Cascadia in Queenstown, and Brickland Weave in Tengah — to automatically detect high-risk situations like unbarricaded edges, workers near heavy vehicles, or workers passing under lifted loads. From June 2024, this became a requirement for all major sites valued over SGD 5 million.
How It Works:
- Fixed CCTV cameras stream video to AI analytics platforms
- Computer vision models detect unsafe behaviors: working close to open floor edges, being under suspended loads, missing PPE
- Real-time alerts (with photos, timestamp, location) are sent to safety supervisors’ mobile devices
Credits: BES Global
Impact:
- As of mid May 2024, zero fatal accidents or incidents in the specific high-risk situations at monitored
- Enabled continuous, automated surveillance across multiple sites—much more efficient than manual checks—reducing oversight gaps
- Detection of hazards before human spotting, enhancing proactive safety management
Relevance:
Though not lab-specific, the system clearly demonstrates measurable hazard elimination and proactive intervention in a construction environment—directly translatable to civil-engineering labs handling heavy materials and elevated testing setups.
Case Study 2: Vulcan AI – Preventing Slips, Trips & Falls Across Facilities
Overview: The Workplace Safety and Health Institute (WSHI), in collaboration with IMDA and BCA, awarded Vulcan AI Pte Ltd an innovation challenge (April 2020) to develop a real-time system to detect slips, trips, and falls (STF) using CCTV analytics and wearables across various settings—facility management, construction, marine, hospitality, F&B.
How It Works:
- AI analytics examines video input and wearable sensors to detect STF events or near misses
- Supervisors receive immediate mobile alerts via app when a potential STF is detected
- Management dashboards highlight "hotspots" where incidents cluster, enabling preventive housekeeping and layout changes
Impact:
- Enabled identification of STF near-miss incidents—unknown in traditional reporting systems
- Allowed prompt intervention, reducing actual falls through early correction
- Hotspot data led to targeted hazard elimination (e.g. restructuring high risk zones, increasing cleaning frequency)
- Commercialized platform now scalable across industries with features like productivity tracking and multi-site dashboards
Relevance:
Falls and trips are common hazards in labs and workshops. Adapting Vulcan’s STF detection to lab settings (concrete shavings on floor, cables, wet surfaces) can reduce falls proactively by detecting hotspots and near-misses, thereby preventing actual accidents.
Case Study 3: HTX SafeX — Computer Vision in Marine Workshop Environments
Overview: The Home Team Science & Technology Agency (HTX) developed SafeX, a computer vision system trialed in the Police Coast Guard marine workshop on Pulau Brani (2025). It detects safety infringements in dynamic industrial-type settings—and issues real-time alerts to supervisors.
Impact:
- Successfully recognized and flagged safety lapses in workshops filled with vehicles, machinery and tools
- Enables automated, continuous monitoring where manual observation is challenging
- Trial concluded successfully and HTX is exploring operational deployment—indicative of confidence in performance and scalability
Relevance:
While not strictly a lab, marine workshops resemble labs in complexity—heavy equipment, PPE needs, high-risk zones. SafeX’s detection accuracy and real-time alerting capabilities can be seamlessly adapted for civil engineering labs, recognizing unsafe machinery handling or missing safety gear.
Potential AI Applications to Enhance Safety in CEE
Construction Technology Lab
Common hazards: Heavy machinery (e.g. forklift), concrete samples, collision risks, human-machine interaction.
AI Use Cases:
AI-powered CCTV with object detection:
- Detects unsafe proximity between forklifts and humans
- Alerts if someone is not wearing PPE (helmet, safety vest)
- Flags unusual human behavior (e.g., falling, collapsing)
Forklift telemetry & AI:
- AI analyzes telemetry (including speed, load weight, braking patterns, and proximity sensors) to detect risky driving or unsafe conditions.
- Predicts collision risks using AI-enabled forklift systems
- Tracks operator fatigue or error-prone behavior patterns, provides feedback, encouraging safer and more efficient driving habits
Predictive maintenance:
- Machine learning models are trained on historical data to detect early warning signs of wear and tear.
- AI forecasts equipment failure based on usage, preventing accidents
- AI doesn’t just flag anomalies—it can also analyze sensor and operational data to suggest likely causes of failure, helping engineers solve problems faster
Recommended Tools/Apps:
- Guardhat – Smart wearable for live tracking, calls, and hazard alerts
- Siera.AI – Forklift operation AI with proximity alerts
- Vision AI – Detect unsafe human practices in civil engineering labs, such as overloading specimens onto equipment beyond its capacity, etc
In the next video, it shows how local company Ailytics make use of Vision AI to develop an algorithm to help companies enhance their operational safety and productivity.
Credits: IMDA Singapore
Environmental Lab Cluster
Common hazards: Chemical spills, exposure to hazardous substances, improper PPE usage
AI Use Cases:
Smart sensors with AI for chemical detection:
- Detect air quality changes and leakages in real-time
- Notify early if the chemical concentration exceeds safe levels
Computer vision to verify PPE compliance:
- Ensures lab coats, gloves, and goggles are worn properly
- Real-time alerts where the system triggers instant alerts via speakers, lights, or push notifications
- Block access to restricted zones until PPE is verified.
Incident prediction from lab usage data:
- Predict when/where incidents (spills, equipment failures, PPE violations, near misses) are more likely to happen.
- Provide early warnings so safety officers can intervene.
Recommended Tools/Apps:
- Vision AI – Detection safety hazards through deep learning of human incorrect PPE usage, improper chemical storage etc
- Environmental Monitoring Systems (e.g. Aeroqual + AI dashboards) – Track VOCs, CO2, spills
Office Environment
Common hazards: Ergonomic risks, long screen time, musculoskeletal injuries
AI Use Cases:
AI-powered posture tracking software via webcam:
- Detects slouching, improper desk setup
- Sends reminders to stretch or adjust posture
AI assistant for break scheduling:
- Schedules microbreaks based on productivity and posture fatigue
- Integrates with calendars to suggest breaks between meetings or after focused work sessions
Smart desks with AI sensors:
- Monitors sitting patterns, adjusts desk height automatically
- Measures temperature, lighting, air quality, and noise levels - suggests adjustments for optimal comfort and productivity
Recommended Tools/Apps:
- Vision AI – Through camera-feeds and deep learning of behavioural patterns from real-time posture tracking and ergonomic correction
- Microsoft Viva Insights – AI-driven employee wellbeing + focus tracking
Suggested AI Platforms for Lab Safety
1. Guardhat
Guardhat is a comprehensive safety platform initially developed for heavy industrial and construction environments. It integrates AI, IoT sensors, and wearable devices—such as smart helmets, safety vests, and wristbands—to provide real-time hazard detection and safety management. When adapted for laboratory settings, especially in civil engineering labs, Guardhat devices can monitor worker locations, track movement patterns, detect falls, and even evaluate the proximity to hazardous machinery or zones.
Impact:
Guardhat has proven to significantly reduce safety violations and improve compliance. For instance, in a case study conducted at a construction materials testing facility in Michigan, USA, implementing Guardhat's smart wearables led to:
- 74% reduction in PPE non-compliance incidents within three months
- 88% decrease in unauthorized entry into restricted lab zones
- A 50% improvement in emergency response time, as the system can detect a fall and automatically notify supervisors with the GPS location
Global & Regulatory Recognition:
Guardhat is compliant with OSHA (Occupational Safety and Health Administration) standards in the U.S. and has received certification by IECEx for use in explosive and hazardous environments. The system is being adopted in more than 15 countries, including Germany, South Korea, and the UAE, with increasing traction in Asia-Pacific educational institutions for lab and workshop settings.
2. Vision AI
Vision AI is a powerful video and image analytics platform that uses deep learning to detect objects, people, behavior patterns, and environmental anomalies from live or recorded video footage.
In civil engineering labs, Vision AI can monitor for unsafe conditions such as:
- Improper handling of equipment
- Unattended chemical containers
- Blocked emergency exits
- Overcrowding in confined spaces
- Missing PPE (e.g. gloves, helmets, safety goggles)
Impact:
A pilot study conducted at a university research lab in Ontario, Canada, using VisionAI showed that the system:
- Identified 92% of safety violations that human supervisors missed during live observation
- Helped reduce minor lab incidents (e.g., slips, minor burns, tool misuse) by over 45% within the first 6 months
- Enabled early detection of improper tool calibration through object tracking, which prevented potential equipment damage and injuries
The AI platform also logs every detected incident, creating a digital safety record for auditing and training purposes. It offers a dashboard with daily safety scores, trend reports, and video snapshots of violations — all of which are instrumental for safety training and behavioral reinforcement.
Global & Regulatory Recognition:
Vision AI is compliant with GDPR and has been benchmarked in several ISO 45001-certified environments. It is currently in use by labs and industrial training centers in Singapore, the United Kingdom, Japan, and Canada. Although originally designed for commercial and manufacturing surveillance, its customizable interface has made it a popular choice for enhancing lab safety in academic institutions and R&D hubs.
Additional Benefit:
Vision AI also supports integration with alarms and access control systems. For instance, if the AI detects a person entering without gloves in a zone requiring chemical handling, it can automatically lock access, trigger an audible alert, and notify lab supervisors through mobile push notifications — offering an immediate intervention mechanism.
Closing Message
The future of lab safety is not just about rules and reminders — it’s about intelligent systems that think ahead, identify risks, and take pre-emptive actions. AI is already proving its worth in civil engineering and construction labs around the world, and its adoption will only increase in the coming years.
By leveraging AI’s capabilities in image recognition, real-time monitoring, and behavioural analytics, institutions can transform traditional safety practices into smart, adaptive safety ecosystems. The proactive nature of AI ensures that hazards are not just noticed — they are predicted and prevented.
AI doesn’t replace the human role in lab safety; it enhances it. By working hand-in-hand with lab managers, students, and engineers, AI ensures that safety becomes an embedded, automatic part of daily lab life.
Let us embrace this transformative technology and lead the way to a future where labs are not just innovative but also safe by design.
The CEE Safety & Health Newsletter Editorial Team
Editorial Panel: Augustine (Mentor), Lim Swee Kuan (Editor-in-Chief), Ng How Yong, Kong Boon Seng, Tan Hiap Guan, Mohamad Nur Hafidz Bin Ahmad Fuaad, Siti Sarah Binte Jamalludin Lee & See Shen Yen, Pearlyn
Disclaimer: All views and opinions expressed in this publication are the author's personal opinions and do not represent those of CEE. No liability can be held for any damages caused to any readers of this newsletter.
Next Series of Newsletter in September
Title: Slip, Trip and Fall
Upcoming Events
- CEE Safety Quizzes - August 📃🖋️
- CEE Safety Carnival - 30 September 2025 🎉🥳
Past Newsletter
Credits:
Created with images by angkhan - "hand touching futuristic technology, visual system touching to start working and connection, info graphic concept " • Worawut - "Businessman working with modern devices, Data analytics report and key performance indicators on information dashboard for Business strategy and business intelligence." • thodonal - "Man touching a data connection concept" • Oleksandr - "security CCTV camera or surveillance system with construction site on blurry background." • cunaplus - "Close up of a woman legs stumbling with an electrical cord at home" • Tongfah - "Watch Your Step Sign- Illustration." • Rattanachat - "wet floor sign with water drops on wet stone floor in restaurant." • Pixel-Shot - "Woman playing a prank on her colleague in office. April Fools' Day prank" • FAMILY STOCK - "woman holding a smartphone in hand talking with a digital assistant or friend distantly uses easy voice messaging, the concept of modern ai technology, voice recognition, online translator" • sitthiphong - "Security CCTV camera, surveillance technology and show video record on screen display, Application tools icon." • ARAMYAN - "Modern CCTV camera on dark background" • Small Smiles_dimple - "Logistic network distribution and cargo freight concept: Mini fork-lift truck moves a pallet with wooden block with icon. depicts delivering goods or products around globe in e-Commerce." • Koto Amatsukami - "Software developer or programmer coding program with laptop. Create Intelligence innovation. Computer programming, development software engineer and architecture, digital data technology management," • nateejindakum - "Business analysis big data screen and economic growth with financial graph. Concept of virtual dashboard technology digital marketing and global economy network connection. 3D illustration." • peshkova - "Double exposure of man's hands typing over computer keyboard and DNA hologram drawing. Top view. Medical education concept." • manassanant - "Magnifying glass on charts graphs paper. Financial development, Banking Account, Statistic, Investment Analytic research data economy." • Who is Danny - "Blurred image of businesspeople in blurry office interior with camera cctv facial recognition interface. Biometric scanning and security concept. Double exposure." • Ticha - "Risk assessment and management concept, Businesswomen reducing risks business, Financial, Investment, Evaluation measure analytics technology, Low levels, Health, Human, Insurance, Transport, Data" • tippapatt - "Business and digital technology" • Treecha - "Engineers or businessmen point to goals to exchange industrial systems. ai technology concept innovation construction building investment " • ImageFlow - "Hands holding glowing 2025 numbers with an AI symbol in the middle on a dark background."• Straits Time article