30 Dec AI reshapes workplace safety practices across global industries

Artificial intelligence is revolutionising how companies approach workplace health and safety, shifting the focus from merely reacting to incidents to proactively preventing risks across various industries around the globe.

EHS (environmental, health, and safety) teams are increasingly adopting AI technologies like computer vision, machine learning, and intelligent video analytics to enhance their risk management efforts. For instance, closed-circuit television systems are no longer just about recording events; they’re now smart enough to detect unsafe behaviours, identify potential hazards, and alert teams before something goes wrong.

Every year, approximately 2.78 million workers lose their lives due to occupational accidents or diseases, according to the International Labour Organization. Additionally, about 374 million workers suffer non-fatal injuries each year. These staggering statistics reveal the shortcomings of safety systems that mainly focus on documenting incidents instead of preventing them in the first place.

A significant challenge organisations face is fragmented data and a lack of real-time visibility, which limits their ability to spot recurring issues and emerging risks across their operations. This is where AI-powered safety systems come into play. They continuously analyse visual and sensor data from worksites, identifying deviations from safe practices within milliseconds, allowing for quicker interventions.

On construction sites, for example, AI can automatically check if workers are wearing the necessary personal protective equipment, ensure vehicles aren’t encroaching on pedestrian areas, and even pinpoint risks of falls or near misses. Some advanced AI tools can even predict when equipment might overheat or fail, helping avert potential accidents.

These intelligent systems are designed to learn and adapt, transforming safety management into a dynamic process rather than just a static checklist. This means AI can understand the context of various situations, distinguishing between routine operations and real dangers.

Governments are starting to acknowledge the invaluable role AI can play in workplace safety. In Singapore, for instance, the Ministry of Manpower has introduced guidelines that advocate for AI-based monitoring of vehicle activities on construction sites. Similarly, Saudi Arabia’s Vision 2030 programme is promoting AI-driven safety technologies for large infrastructure projects.

The International Labour Organization has also emphasised the importance of AI in workplace safety, particularly during World Day for Safety and Health at Work 2025, showcasing a growing international consensus on the matter. Construction, which regulatory bodies like OSHA deem one of the most hazardous industries, recorded the highest number of workplace fatalities, mainly due to falls and transport-related incidents.

On a significant infrastructure project in Singapore with over 12,000 workers, AI-enabled safety systems have been set up to identify hazard zones like open edges, confined spaces, and areas with heavy machinery. Now, supervisors receive real-time alerts when risks rise due to congestion, poor lighting, or equipment movements.

In the mining sector, which tragically accounts for a high number of workplace deaths, AI is being deployed to tackle risks associated with confined spaces, gas exposure, and equipment failure. Companies like Rio Tinto and BHP are using AI-driven monitoring and digital twin technologies to assess safety conditions and support predictive maintenance.

Even in manufacturing, firms are turning to AI analytics to minimise machinery-related incidents. Companies like Siemens and General Electric are utilising machine learning models to monitor data on vibrations, temperature, and acoustics, allowing them to identify early signs of equipment failure and schedule timely maintenance to reduce injury risks.

While the potential of AI in safety management is immense, it heavily relies on the quality of data and ethical usage. Poor or biased data could lead to false alarms or missed hazards, and thus many organisations are adopting human-in-the-loop approaches where safety professionals review AI alerts.

Privacy concerns also loom large. Continuous monitoring prompts questions regarding surveillance and workers’ rights. In response, some companies are implementing privacy-preserving AI that anonymises workers while still detecting unsafe behaviours.

However, implementing these systems can be challenging, especially for small and medium-sized enterprises lacking advanced infrastructure. Integrating AI tools with existing EHS systems can also pose technical and organisational hurdles. Additionally, safety professionals need new skills to effectively interpret data and manage the insights that AI provides.

Ultimately, AI is not set to replace the core principles of workplace health and safety. Instead, it’s reshaping the way we apply those principles for a safer work environment.