AI's Role in Evolving Workplace Safety: NSC's Vision for a Zero-Incident Future

In today's data-driven world, AI isn't just on the map - it's the fast lane. From mining critical insights and tracking regulatory shifts to forecasting risk, monitoring employee health and safety metrics, and streamlining reporting, AI is steering the future of EHS, sustainability, and compliance. But as professionals take the wheel, what roadblocks and green lights lie ahead on the journey? This series of articles will help you navigate the future of AI in EHS.

There have been significant strides made in recent years to improve workplace safety and reduce or eliminate workplace fatalities, injuries, and illnesses. Now, as companies seek new ways to reduce risk and better protect workers, safety experts are turning to artificial intelligence (AI) as a potential game-changer - and asking a critical question: Can AI help reverse the trend and move workplaces closer to a zero-incident future?

The National Safety Council (NSC), a nonprofit dedicated to eliminating preventable workplace injuries and deaths, is actively exploring that very question. As more industries digitize operations and generate real-time data, the potential to apply AI within EHS (environmental, health, and safety) programs is greater than ever.

To understand how AI is shaping safety strategies - and how EHS professionals are adapting - 3E spoke with Jay Vietas, NSC's senior director of research and an expert on emerging technologies, including AI's role in worker health and safety.

AI's Drive Toward Zero-Incident Workplaces

Achieving zero-incident workplaces is one of the NSC's most ambitious safety goals and an area where Vietas believes AI will play a critical role.

“AI can help organizations move closer to zero-incident environments in several ways. First, it can help automate higher-risk work, removing workers from dangerous situations. Second, AI can quickly identify high-risk behaviors and alert supervisors or workers before incidents occur,” Vietas explained, adding that safety professionals can use advanced analytics and AI to better understand risks and make more informed decisions about their safety programs.

One of the key initiatives supporting this vision is the NSC's Work to Zero program, a strategy aimed at eliminating workplace fatalities by 2050. The program emphasizes accelerating the adoption of advanced technologies such as AI, sensors, drones, wearables, computer vision, virtual reality, and augmented reality.

“Computer vision products are being used to identify hazards and alert workers and [supervisors] before incidents occur,” Vietas said. “These tools can be programmed to recognize potential safety issues in real time and provide immediate warnings.”

In addition to large-scale initiatives like Work to Zero, the NSC is also embracing AI to strengthen its internal operations. Vietas noted that the organization has begun using large language models (LLMs) to increase efficiency and improve the quality of its day-to-day work. LLMs, such as ChatGPT, are AI models capable of a wide range of tasks, including answering questions, summarizing text, translating languages, and even creating content.

“We continue to explore technical solutions, including AI, to improve the quality and timeliness of our products and services. Our goal is to advance our safety mission by using these tools to serve our members better,” Vietas said.

AI's Emerging Role in Chemical Safety

Beyond physical safety, AI is also beginning to reshape chemical risk management - another area where speed and precision are critical. According to Vietas, AI tools are already being used to monitor hazardous environments in real time, detecting unsafe chemical levels and alerting workers or supervisors before exposures become dangerous. When paired with sensors, he said, these systems can trigger automatic warnings, enabling faster responses and potentially preventing long-term health issues.

Beyond real-time monitoring, researchers see even broader applications on the horizon, particularly in how AI could support emergency medical response. Vietas highlighted the ongoing work of researchers at the NICEATM, the National Toxicology Program's Interagency Center for the Evaluation of Alternative Toxicological Methods. There, he said, scientists are using AI and machine learning to predict chemical toxicity, reduce the need for animal testing, and support regulatory decisions. Tools like STopTox (Systemic and Topical Chemical Toxicity) and CATMoS (Collaborative Acute Toxicity Modeling Suite) help forecast how chemicals might affect the human body. Others, like SARA-ICE (Skin Sensitization Risk Assessment – Integrated Chemical Environment) and the DASS App (Defined Approaches for Skin Sensitization), assist in evaluating skin sensitization risks. These platforms draw on large datasets and apply complex modeling to make chemical risk assessments faster and more accurate.

Although the use of generative AI in real-world emergency settings is still in its early stages, Vietas believes its potential is significant.

“First responders could likely benefit from generative AI to improve patient treatment in emergency situations,” he explained.

AI in Environmental Compliance: Opportunities and Roadblocks

AI is gaining traction in safety and chemical monitoring. However, its use in environmental compliance is still evolving. Vietas noted growing interest in using AI to track emissions, monitor environmental conditions, and flag compliance risks, but measurable, industry-wide results are still limited, he said.

“We are still in the early stages of AI adoption for environmental compliance and monitoring,” he said. “While we expect to see improvements as these tools mature, consistent and measurable results aren't yet widely visible across the industry.”

He emphasized that the challenges extend beyond technology to strategic concerns. Common barriers include high implementation costs, limited awareness of available AI solutions, and insufficient access to quality data for training models. Without clear evidence of tangible results or a strong return on investment, he said, many organizations remain hesitant to commit.

Additionally, Vietas emphasized that successfully integrating AI into EHS workflows often demands broader organizational changes.

“Some companies aren't ready to implement the changes required for these tools to work effectively,” Vietas explained. “And data protection and privacy concerns also need to be addressed before these systems can be fully trusted and deployed.”

Despite these hurdles, Vietas remains optimistic that as infrastructure improves and regulatory demands intensify, AI will shift EHS teams from reactive compliance to proactive performance.

“The core role of EHS managers will largely remain the same,” Vietas said. “But they'll need to understand both the benefits and risks of AI systems, and work closely with engineers and IT professionals to safely implement and manage these tools across their organizations.”