Using AI to Enforce Safety Compliance in Manufacturing

It’s no secret that injuries are a major concern in manufacturing, and human errors associated with these injuries cost organizations billions every year. In every environment, safety equipment is required to protect workers from injuries, but researchers found that many employees do not wear protective gear when necessary. Leveraging artificial intelligence (AI), the manufacturing industry can protect workers better and ensure that all employees wear gear to stay compliant with Personal Protective Equipment (PPE) protocols to avoid costly penalties.


Non-Compliant Work-Related Injuries

According to the National Safety Council, the total cost of work injuries in 2019 was $171.0 billion in all industries. Manufacturing accounted for 15.0 percent of all private industry nonfatal injuries in 2019, according to the Employer-Reported Workplace Injuries and Illnesses – 2019.

One of the major causes of injury is a non-compliance to wearing safety equipment according to a survey released by Kimberly-Clark Professional, which reported that 89 percent of safety professionals said they had observed workers not wearing safety equipment when they should have been. Twenty-nine percent said this had happened on numerous occasions. Such a high rate of noncompliance with PPE protocols presents a threat to worker health and safety, which may cause serious injury or even death.

Worker safety and hefty penalties for non-compliance are a major concern for organizations, making it even more critical for training and monitoring. Even though non-compliance harms employees and can be fatal, many of them do not understand the consequences and potential for injury should they not wear PPE gear. For the organization, they lose an employee and also bear the responsibility of paying for workman’s compensation. Insurance covers payments for injuries, but it increases future premium costs.


Protecting Workers and Staying Compliant with AI Monitoring

Staying compliant with PPE protocols requires employee training and management. For the organization as a whole to remain safe, employees and management must train coworkers and correct any mistakes. It’s impossible to completely eliminate risk of human error, but organizations can implement strategies that will greatly improve conditions and reduce risk. Artificial intelligence can help with your organization’s safety mission and keep it compliant with PPE protocols and requirements.


Manufacturing organizations already have cameras set up in strategic positions, and AI solutions can integrate with pre-existing monitoring including cameras to detect employees without protective gear. The AI algorithms can analyze employees on-camera and determine if they are wearing the right gear (e.g., helmets or goggles). Algorithms in AI are programmed to “learn,” which means that they can evolve over time as they continue to discover compliance issues. As more data is fed to the algorithms, management receives fewer false positives as the AI system gets better at recognizing employees missing their PPE gear.


As AI adapts to your monitoring system and images, it can differentiate between different types of gear. For example, it can determine if a worker is wearing a PPE hard hat or a standard baseball cap. This benefit will stop employees that come up with ways to bypass the monitoring system to avoid wearing gear. It will also stop false positives from alerting management and causing too much overhead. With any properly tuned AI system, it must be trusted or alert notifications lose their value. When management does not trust the system, they might ignore important alerts thinking that another false positive was triggered.


The way machine learning (ML) algorithms work to identify non-compliance (i.e. anomalies) is in the way it digests data. The ML solution processes real-time video streams captured by cameras. Video streams are sliced frame-by-frame; every frame is treated as a separate image analyzed by the algorithm and may be processed on the edge to ensure faster ML model response using AWS Panorama Appliance or Nvidia Jetson devices. The algorithm itself is trained to identify PPE violations (i.e. a construction worker not wearing a high-visibility vest) and ping the alerting part of the system to send a notification to the system’s front end (e.g., a smartphone application). Management can then view the alert and make corrective action, to gradually adjust and finetune the system, so that the employee can be corrected and trained. It also helps security advisors determine better ways to train employees so that they understand the importance of compliance and work safety.



AI gives more resources back to workers, especially security personnel. Where supervision may not help, proper training does — and AI enables security officers to develop more advanced training programs and spend more time training workers. Also, consider AI-powered worker safety solutions a habit-building and support tool: notifications on violations may not necessarily cause penalties; rather, they remind workers about what they already know — be vigilant, stay safe, and follow regulations. AI just helps reinforce the safety and security message through technology.


Rinat Akhmetov is the ML Solution Architect at Provectus. With a solid practical background in Machine Learning (especially in Computer Vision), Rinat is a nerd, data enthusiast, software engineer, and workaholic whose second biggest passion is programming. At Provectus, Rinat is in charge of the discovery and proof of concept phases, and leads the execution of complex AI projects.



Comments (0)

This post does not have any comments. Be the first to leave a comment below.

Post A Comment

You must be logged in before you can post a comment. Login now.

Featured Product

BigRep ONE: Large-Scale 3D Printing

BigRep ONE: Large-Scale 3D Printing

The BigRep ONE is an award-winning, large-format 3D printer at an accessible price point. With over 350 systems installed worldwide, it's a trusted solution for prototyping and production by designers, innovators, and manufacturers alike. Featuring a massive one-cubic-meter build volume, the fast and reliable ONE brings your designs to life in full scale.