For as long as there has been work to be done, there have been hazards associated with doing it. Whether these hazards take the form of falls, interactions with vehicles and heavy equipment, or exposure to chemicals or other substances, a critical challenge of modern industrial organizations across the globe is ensuring the safety of workers and individuals living near industrial sites and workplaces. In years past, this has been a largely reactive pursuit, i.e., analyzing the causes of accidents, toxic releases, etc. only after they have taken place. But with the emergence of artificial intelligence (AI) in recent decades, industry leaders are now employing this technology in a wide range of ways to further minimize health, safety, and environmental (HSE) risks, enhancing worker safety while maximizing productivity.
The range and impact of hazards is immense and growing
A recent report indicates that approximately 5,283 workers were killed in on-the-job accidents in just one annual reporting cycle. Leading causes of these incidents include transportation (driving or pedestrian-vehicle accidents), slips and falls, workplace violence, and exposure to harmful substances, electricity, or temperature. Especially hazardous industries include construction, agriculture and forestry, transportation and warehousing, and manufacturing. In addition to fatal injuries, there are, of course, far more non-fatal injuries, not only in the above industries but in some less expected areas like healthcare, package delivery, and veterinary services. Recent OSHA accident statistics indicate that there were 2.6Mn workplace injuries in the US annually or 2.4 incidents per 100 FTE. That’s a slight decrease from the preceding year, but still a major challenge for industrial firms, not only because of the human cost but also the downtime, productivity losses, medical expenses, and potential asset damage frequently associated with these events.
AI plays a significant role in reducing accidents
Despite the generally negative tone of industrial accident statistics, there is at least one positive aspect of these events, one that lends itself well to the use of AI in mitigating hazards. This is the fact that many hazardous conditions can be recognized visually. Add to this the fact that most industrial facilities are already well equipped with significant camera infrastructure, and you have an environment that’s ready–made to take proactive steps toward reducing or eliminating workplace accidents and hazards. Specific hazards that are proactively discernible with cameras include (but are by no means limited to):
- Personal protective equipment (PPE) compliance: hard hats, safety glasses, steel-toed boots, reflective vests, gloves, safety harnesses, etc.
- Proximity/near misses of workers to vehicles or heavy equipment
- Workers beneath suspended loads or at-height workers/tools
- Use of inappropriate/unsuitable tools for the job
- Fire/smoke alerts, particularly in unattended areas (including flagging of overheating equipment using infrared cameras)
- Gaseous or fluid chemical releases/spills
- Unauthorized/unqualified workers in controlled areas (by scanning of faces or ID badges and comparing these against personnel qualification databases)
- Security perimeter breaches
- Weapon detection, active shooter events
The list of such use cases is quite literally endless, ranging in severity from mere nuisances to fatal outcomes. By training AI visual systems to identify these use cases, numerous mitigating actions can be implemented:
- Send proactive alerts to workers and/or supervisors concerning PPE compliance, entry into hazardous areas, etc.
- Turn on area danger signals (horns, klaxons, lights) alerting workers to unsafe situations/conditions
- Autonomously/remotely shut down equipment if worker approaches too closely
- Summon emergency responders if necessary
In addition to providing proactive, actionable alerts that mitigate impending safety situations, AI also supports valuable periodic reporting that helps managers track hazard trends and identify the most serious risks, enabling implementation of subsequent safety-enhancing capabilities, e.g., railings, harnesses, high-visibility paint, non-skid matting, etc. Such reporting is also valuable in educating employees to the hazards that exist in their workplaces.
But it’s not just about cameras
In addition to using camera imagery to spot and alert on impending hazards in industrial workplaces, AI-enabled tools like digital twins can perform analysis of workplaces before employees even show up to work, identifying potential human-vehicle pinch points, trip hazards, and other risks workers should know about before starting their shifts. And by equipping workers with wearables and biometric devices, health issues like heart rate, respiration, physical positioning, etc. can identify incipient conditions like fatigue, muscle strain, and poor attention before such conditions can lead to incidents. These capabilities, combined with analysis of pending weather, excessive heat, and other environmental conditions, can be used in risk mapping, identifying hazards proactively, and helping mitigate risks to workers. Using these capabilities in trending and history analysis can even highlight especially hazardous days/times for workers, raising awareness and caution.
Cultural concerns should be considered prior to implementation
An important consideration of these advanced capabilities is the potential concern of workers with privacy and worker/supervisor relations, an issue of particular sensitivity in unionized workplaces. To the extent that HSE mitigation is sometimes reliant upon identifying specific individuals (and their level of training, seniority, authorization, etc.), some workers may be uncomfortable with such processes. AI’s ability to mask faces and other unique identifiers can help to address these concerns, but it is nonetheless important to address such issues before implementing AI-powered HSE systems.
Conclusion: AI is just a tool; results derive from culture
Avathon’s Autonomy Platform enables managers to automatically monitor HSE status, assess risks, and deliver real-time notifications to team members that reduce or eliminate incidents. The platform can also autonomously shut down equipment, immediately mitigating risks without relying on employee actions. Our Visual AI technology autonomously sees and interprets employee actions and environmental conditions wherever cameras are in place, creating opportunities to enhance safety and increasing the value of a company’s existing camera infrastructure.
Organizations have the flexibility to configure the Autonomy Platform to best address their specific process, environmental, and infrastructure challenges, whether at a construction site, manufacturing plant, or oil platform. By generating autonomous alerts for HSE compliance failures and other risks, safety is enhanced, costs reduced, and company reputation improved.
Continuing workplace risks and worker accidents do not have to be accepted as the normal state of affairs. Advances in computer vision and visual AI technology have achieved a level of maturity and adoption that allows these systems to deliver genuine value to enterprises and their employees. The Avathon Autonomy platform minimizes time-to-results for computer vision technology, delivering sustainable HSE benefits across a wide range of industries and operational processes.
Visit our site to learn more about Avathon’s Visual AI and HSE solutions.
