“The integration of Artificial Intelligence (AI) into business processes has created a paradigm shift in how organizations approach decision-making, efficiency, and now, employee safety. In the realm of Environmental, Health and Safety (EHS) performance, AI’s potential is especially transformative, due to its ability to deliver deep insights.”

–Chet Brandon & Microsoft Copilot

The American Society of Safety Professionals AI Task Force: The Role of AI in the Future of Work & Safety

I am excited by a new initiative from the The American Society of Safety Professionals. The organization is forming a time-focused task force to explore the impact of Artificial Intelligence (AI) on safety practices across industries. They are seeking knowledgeable safety professionals to volunteer as task force members. This role will involve collaborative work to assess AI-driven strategy, safety products and services, develop case studies, and contribute to the evolution of ASSP’s education and training offerings. Review research conducted by ASSP’s Board Working Group on AI applications in safety, reviewing emerging trends, challenges, and opportunities. Review and prioritize research outcomes to identify sequencing and focus of the task force efforts. I have applied and hope to be selected for the Task Force. If you are an ASSP member, you to can apply at this link: (deadline past).

January 2025 Update: I was selected for this task force. The task force has begun its work and we’ll have feedback for the ASSP Board of Directors in time for their March meeting.

Update on the Final Work Product of ASSP‘s AI Task Force (Q2 2025)

The task force did complete its work, and that was combined with member input, and direction from ASSP‘s board, to develop 5 insights and observations regarding AI and Occupational Safety and Health:

• Understanding impact on professions and jobs. AI will likely change job roles and require change management, including identifying necessary skills and providing training. However, use of AI does not reduce the role of safety professionals. Human interaction is essential as organizations cannot rely solely on AI for tasks that require human expertise and decision-making. 
• Improving awareness. Many OSH professionals report they lack knowledge about AI and its implications, creating a need for increased education and awareness. 
• Providing training and guidance. OSH professionals also indicate a need for training on intermediate-level AI, best practices and ethical implementation. Other identified needs include guidance on verifying information and improving questioning techniques, as well as resources such as articles, case studies and benchmarking. 
• Defining ethical implications. Concerns with the use of AI include biases and the impact on professional obligations and certifications. Members and the OSH community will look to ASSP for guidance on ethical use and implementation of AI in OSH. 
• Protecting privacy and data. AI raises concerns about privacy and data protection, and addressing these issues will continue to be critical to implementation and acceptance.

The final output that the task force participated in can be found on ASSP‘s website at this link:

https://www.assp.org/about/artificial-intelligence—safety

As part of this initiative ASSP is calling for case studies demonstrating successful application of AI technology in the OSH, Sustainability and Environmental fields. ASSP also has some informative free training resources available at the above link.

Purpose of this Article

I have created the article below to capture my thoughts on the effort and identify how I could contribute. This is a thought exercise that I conducted with the help of Microsoft’s AI technology, named Copilot. I created it by inputting general and specific ideas I wanted to explore further and then adjusting my query in Copilot until it produced the answer I felt best reflected my idea and intentions. Copilot and I co-authored most of this post. ChatGPT contributed some as well and it’s interesting to see the differences in answers between the two. I’m interested in what you think of it. By the way, Copilot created the picture above from this description: Create an image of AI helping an industrial manufacturing worker.

The Origin of My Interest

The integration of Artificial Intelligence (AI) into business processes has created a paradigm shift in how organizations approach decision-making, efficiency, and now, employee safety. The ability to process vast amounts of data swiftly and accurately has provided leaders with unprecedented insights into their operations. In the realm of Safety and Environmental Health and Safety (EHS) performance, AI’s potential is especially transformative, due to its ability to focus leaders on improvement actions identified by these deep insights. This article delves into how AI enhances decision-making, improves safety management, and addresses ethical considerations, drawing on practical experiences and collaborations, including insights from my partnership with Benchmark Gensuite since 2019. But my interest in the area actually started in 2014. I have a blog post on this site from that time: Is EHS on the Verge of Disruption?

The Power of AI in Decision-Making

AI’s impact on decision-making is profound. Traditional decision-making processes often rely on historical data and human intuition, which can be biased or slow. AI, however, can analyze large datasets in real-time, identify patterns, and provide predictive insights. This capability is crucial for EHS management, where timely and accurate decisions can prevent accidents and save lives.

For example, natural language processing (NLP) can be used to analyze large volumes of text records to identify trends and inherent risks. By examining incident reports, maintenance logs, and inspection records, AI can pinpoint recurring issues and potential hazards that might be overlooked by human analysts. Similarly, deep learning algorithms can monitor real-time video feeds from the work floor, identifying at-risk behaviors of employees and flagging them for immediate action.

Practical Applications: Partnership with Benchmark Gensuite

Since 2019, I have been actively involved with Benchmark Gensuite, a leading provider of online EHS Digital Management Systems. This collaboration aims to enhance EHS management using AI technologies such as machine learning, neural networks, and NLP. By integrating these AI capabilities into Benchmark Gensuite’s commercial management systems, we have created more robust and efficient tools for managing EHS compliance. Today, this system serves over 1.5 million users worldwide.

One of the key initiatives in this collaboration is the development of AI-driven risk assessment models. These models use machine learning algorithms to analyze data from various sources, such as safety audits, incident reports, and environmental sensors. The insights generated by these models help organizations proactively address potential risks, reducing the likelihood of accidents and improving overall safety performance.

Envisioning the Future: AI in the Safety Profession

The role of AI in evolving the safety profession cannot be overstated. AI is not just a tool for improving existing processes; it is a catalyst for innovation and transformation. By leveraging AI, organizations can drive standardization in safety practices, enhance process efficiency, and ensure compliance with regulatory requirements, even in resource-constrained environments.

A Discussion of AI Processes that Can Be Applied in the Work Setting

Artificial Intelligence has the potential to revolutionize employee safety in manufacturing settings by creating smarter, proactive systems that minimize risks and improve operational efficiency. By leveraging AI’s capabilities to monitor, analyze, and predict workplace conditions, manufacturers can transform their operating environments into safer, more responsive ecosystems. AI-enhanced systems can detect hazards in real time, optimize workflows to reduce human exposure to dangerous tasks, and provide data-driven insights for continuous safety improvements. These technologies not only reduce accidents and injuries but also boost productivity by ensuring a safe and compliant work environment. The methods listed below highlight how AI can be applied across various facets of safety in manufacturing, illustrating its transformative impact on employee well-being and operational excellence.

AI processes are the foundational approaches and techniques that enable artificial intelligence systems to simulate human-like intelligence by processing data, learning from it, making decisions, and performing tasks. A crucial aspect of these methods is their ability to understand context, which allows AI systems to interpret information more effectively and respond appropriately to different situations. For instance, machine learning enables AI to identify patterns and adapt over time, while natural language processing not only understands words but also the intent and nuances behind them. Similarly, deep learning models can analyze complex relationships in data, considering contextual factors to refine predictions or decisions. This capability to grasp context ensures that AI systems can deliver relevant and accurate outcomes, whether they’re diagnosing a problem, providing recommendations, or interacting naturally with humans. Contextual understanding makes AI systems smarter and more versatile, paving the way for their effective application in dynamic environments like healthcare, manufacturing, and customer service. A comprehensive discussion of most of the currently in use and near future processes is below.

Machine Learning (ML)

ML enables predictive and adaptive systems that enhance workplace safety in manufacturing. By analyzing historical accident data, supervised learning can predict high-risk scenarios, while unsupervised learning can identify hidden patterns leading to unsafe conditions. For instance, ML models could monitor machinery data to forecast failures that might cause accidents, prompting preventive maintenance and reducing risks.

Deep Learning

Deep learning processes complex data like video feeds to monitor safety compliance. Convolutional Neural Networks (CNNs) can detect unsafe behaviors, such as workers not wearing protective gear, and alert supervisors in real time. For example, a deep learning-based system could analyze CCTV footage to ensure compliance with safety protocols like hard hat usage or proper equipment handling.

Natural Language Processing (NLP)

NLP interprets and processes text or speech data, improving communication and hazard reporting. Workers can use voice-activated systems powered by NLP to report safety incidents instantly, ensuring quicker response times. For example, an NLP-driven chatbot might allow employees to describe hazards in natural language, which the system categorizes and prioritizes for resolution.

How Large Language Models Relate to NLP 

Large Language Models (LLMs) are a powerful and advanced application of Natural Language Processing (NLP) — in fact, they represent one of the most significant breakthroughs in the field.

NLP is the broader domain focused on enabling machines to understand, interpret, generate, and respond to human language. It includes tasks like text classification, sentiment analysis, machine translation, question answering, and more.

LLMs, such as OpenAI’s GPT models, are built using deep learning techniques—especially transformer architectures—and are trained on massive amounts of text data to learn the statistical patterns of language. As a result, LLMs can perform a wide variety of NLP tasks without being explicitly programmed for each one. What sets LLMs apart is their ability to generate coherent, contextually relevant, and fluent language based on prompts, making them adaptable to many use cases including chatbots, document summarization, content creation, and intelligent virtual assistants (IVAs).

Computer Vision

Computer vision automates safety inspections and monitors real-time conditions. It can detect hazards, such as spills or obstructions, and notify employees or halt operations. For instance, a vision system could monitor manufacturing lines to ensure workers maintain safe distances from dangerous machinery, reducing the likelihood of accidents.

Expert Systems

Expert systems use predefined rules to assess and mitigate risks. In manufacturing, these systems can evaluate compliance with safety regulations or suggest corrective actions after identifying risks. For example, an expert system could review workplace conditions against OSHA standards and recommend safety improvements in real time.

Evolutionary Algorithms

These algorithms solve optimization problems related to safety planning. For example, genetic algorithms could optimize factory layouts to minimize worker exposure to hazardous zones. A manufacturing facility might use such algorithms to design workflows that reduce interactions between humans and heavy machinery, lowering accident rates.

Fuzzy Logic

Fuzzy logic handles uncertain or imprecise data, making it useful for real-time safety decisions. It can assess conditions such as excessive vibration or temperature changes in machinery that might signal potential hazards. For example, a fuzzy logic system could decide when to trigger alarms for borderline unsafe conditions, allowing timely interventions.

Robotics

AI-powered robots improve safety by handling dangerous tasks, such as working with hazardous chemicals or performing repetitive, injury-prone activities. For example, collaborative robots (cobots) in manufacturing can work alongside humans, taking over heavy lifting tasks to reduce worker fatigue and prevent musculoskeletal injuries.

Knowledge Representation and Reasoning

This method structures and utilizes safety-related knowledge for decision-making. A knowledge graph could integrate incident data, equipment logs, and safety regulations, enabling intelligent systems to recommend actions. For example, a system might cross-reference equipment failures with historical accidents to highlight areas needing immediate attention.

Hybrid AI

Hybrid AI combines multiple techniques for comprehensive safety monitoring. For instance, integrating computer vision with ML could analyze both visual and sensor data to detect hazards more accurately. A manufacturing facility might deploy a hybrid system to ensure workers follow safety protocols while machines operate within safe parameters.

Probabilistic Methods

These methods assess risks and uncertainties to prioritize safety interventions. For example, Bayesian Networks could model the likelihood of an accident based on current conditions like machinery status and worker fatigue. A facility could use such models to predict and address high-risk situations before accidents occur.

Cognitive Computing

Cognitive systems simulate human reasoning to assist in safety decision-making. For example, a cognitive assistant might analyze a combination of weather conditions, machinery data, and worker schedules to recommend safety measures, such as delaying operations in extreme conditions.

Complex Quantitative Problem Solving

AI tools like ChatGPT can serve as powerful co-pilots for EHS professionals by handling complex quantitative tasks such as chemical exposure assessments and energetic chemical reaction analysis. They can quickly perform and explain calculations like time-weighted averages, short-term exposure limits, and adiabatic temperature rises, while also structuring hazard analysis frameworks and identifying errors in assumptions or unit conversions. These capabilities make AI valuable for speeding up routine computations, scaling exposure assessments across large datasets, and translating technical findings into clear language for broader audiences. However, accuracy depends on high-quality input data, and AI cannot replace specialized simulation software, experimental validation, or expert engineering judgment—meaning it is best applied as a decision-support tool that enhances, rather than replaces, professional oversight in safety-critical EHS work.

Quantum AI

Though emerging, quantum AI offers potential in optimizing safety-critical systems. For instance, a manufacturing plant could use quantum algorithms to optimize emergency evacuation routes considering real-time hazards, minimizing risks during accidents.

Multi-Agent Systems

Multi-agent systems involve AI agents collaborating to monitor and enforce safety. For instance, agents could simulate workflows to identify potential hazards or coordinate responses to emergencies. A manufacturing plant might use these systems to ensure multiple safety processes work in tandem, such as fire suppression and evacuation protocols.

These AI methods, tailored to specific safety challenges, can significantly reduce risks, enhance compliance, and ensure a safer environment for manufacturing workers.

Intelligent Virtual Assistant

An Intelligent Virtual Assistant (IVA) is highly relevant to occupational health and safety (OHS) professionals as it serves as a powerful tool for improving workplace safety and compliance. IVAs can assist in real-time hazard identification by monitoring data from sensors, cameras, and IoT devices, alerting professionals to potential risks. They can streamline safety reporting by enabling workers to log incidents or unsafe conditions through voice or text commands, ensuring timely documentation and response. Additionally, IVAs can deliver personalized training, answer safety-related questions, and ensure workers are informed about safety protocols. For OHS professionals, IVAs provide data insights and trend analysis, helping them identify recurring issues and implement preventive measures. By automating routine tasks and enabling better communication, IVAs empower OHS professionals to focus on strategic initiatives that enhance workplace safety and protect employees.

Empowering Manufacturing Through AI-Driven Sensemaking and Environmental Integration

The transformative potential of AI in revolutionizing manufacturing operations and the safety of those operations will be significantly amplified when integrated with a network of diverse input systems. These systems could include advanced sensors, video feeds, IoT devices, and real-time data streams, enabling the AI to achieve a sophisticated level of sensemaking. Sensemaking is the process of interpreting and understanding complex or ambiguous information to create meaningful insights and guide decision-making. By comprehensively interpreting the physical and contextual environment in which it operates, the AI can offer nuanced insights, predict challenges, and identify opportunities for optimization.

AI systems can perform a significantly deeper level of sensemaking beyond human capability due to their ability to process and integrate vast amounts of data from diverse sources at incredible speed and precision. Unlike humans, who are limited by cognitive bandwidth and biases, AI can analyze complex, multidimensional datasets simultaneously, identifying patterns, correlations, and trends that might remain imperceptible to human observers.

Additionally, AI systems can leverage advanced algorithms, such as machine learning and deep learning, to continuously refine their understanding and adapt to changing environments. They can integrate real-time data from the diverse range of input devices discussed above, and digital models, providing a comprehensive and dynamic understanding of a situation. These systems can also operate 24/7, maintaining consistency and avoiding fatigue or errors caused by human limitations.

Furthermore, AI can simulate and predict outcomes using historical data, enabling proactive decision-making. By combining cognitive computing with predictive analytics, AI can anticipate future states, assess risks, and recommend or implement optimal solutions, achieving a depth and breadth of sensemaking far beyond human capacity.

To maximize the benefits of such an AI ecosystem, its functionality should extend beyond data analysis and decision-making by connecting to an array of output devices. These devices could range from robotic actuators and autonomous machinery to interactive dashboards and augmented reality interfaces. This integration would allow the AI not only to communicate actionable insights but also to directly influence and manipulate the physical environment. Such capabilities can result in tangible enhancements in operational efficiency, proactive risk mitigation, and adaptive responses to dynamic manufacturing conditions, ultimately fostering safer, smarter, and more resilient production environments.

An advanced level of sensemaking in AI systems can revolutionize employee protection in manufacturing by enabling a holistic, proactive, and dynamic approach to workplace safety. Beyond mitigating risks, these systems free human workers to focus on areas where their unique capabilities, such as creativity, intuition, and complex problem-solving, are most effective. Here’s an expanded view:

1. Hazard Identification and Risk Prediction: AI systems can process inputs from multiple sources, such as IoT sensors, thermal imaging, and acoustic analysis, to detect anomalies and predict hazards with unparalleled accuracy. For example, AI can anticipate equipment failure by analyzing vibrations or thermal data, preventing accidents caused by malfunctions. This predictive capability surpasses human limitations, offering a constant safeguard against emerging risks.
2. Real-Time Monitoring and Alerts: With its ability to analyze vast datasets instantly, AI can monitor the workplace in real time, identifying unsafe conditions or behaviors such as machine misuse, slip hazards, or chemical leaks. Immediate alerts and automated interventions can halt potentially dangerous activities before harm occurs, significantly reducing reaction time compared to human responses.
3. Personalized Safety Interventions: AI can assess individual worker conditions using wearable devices that track heart rate, temperature, and movement patterns. For instance, it can detect early signs of fatigue, dehydration, or stress and recommend personalized measures, such as breaks or hydration, ensuring workers remain physically and mentally fit for their tasks.
4. Automation of High-Risk Tasks: Dangerous tasks, such as heavy lifting, handling hazardous substances, or working in extreme environments, can be assigned to AI-controlled robotic systems. This reduces direct human exposure to risks while maintaining productivity and precision.
5. Enhanced Training and Simulation: AI-driven virtual reality (VR) and augmented reality (AR) tools can immerse employees in realistic training scenarios, teaching them how to handle emergencies or follow complex safety protocols. These adaptive learning systems tailor content to individual needs, improving comprehension and retention.
6. Continuous Safety Improvement: AI systems can analyze historical incident data to identify trends, root causes, and potential improvements. These insights enable the implementation of evolving safety measures, ensuring a dynamic approach to workplace security.
7. Dynamic Environmental Adaptation: AI can instantly recalibrate safety protocols in response to changes in the work environment, such as the introduction of new machinery or workflow alterations. This ensures ongoing alignment with safety standards without human intervention.

By taking over routine, high-risk, and data-intensive tasks, AI allows human workers to focus on areas where their unique strengths shine:

1. Creative Problem-Solving: Humans can address complex challenges requiring innovation, such as optimizing manufacturing processes or designing new products, while AI handles routine operational decisions.
2. Interpersonal Communication: Employees can dedicate more time to collaborative roles, such as team leadership, mentorship, and customer engagement, fostering a culture of cooperation and continuous improvement.
3. Strategic Oversight: Workers can concentrate on high-level decision-making, interpreting AI-provided insights within the broader organizational context, and aligning them with business goals.
4. Ethical and Social Considerations: Humans are uniquely suited to consider ethical implications, cultural sensitivities, and emotional nuances that AI cannot fully comprehend. This ensures that safety and operational strategies are implemented humanely and inclusively.
5. Adaptability and Contextual Understanding: While AI excels in pattern recognition and prediction, humans bring context, intuition, and the ability to adapt to unforeseen circumstances, ensuring balanced decision-making.

By leveraging AI’s advanced sensemaking capabilities to enhance safety and offload repetitive or hazardous tasks, manufacturing environments can empower employees to excel in roles that require creativity, empathy, and strategic thinking. This human-AI collaboration not only creates safer workplaces but also drives innovation and workforce satisfaction.

A related by attribute of AI functionality is Situational intelligence. Situational intelligence refers to the ability to perceive, comprehend, and respond effectively to dynamic and context-specific circumstances. It involves the integration of real-time data, historical knowledge, and contextual awareness to make informed decisions or take appropriate actions in a given situation. This AI attribute improves occupational safety in manufacturing by enabling real-time monitoring and proactive risk mitigation. IoT sensors, wearable technology, and connected systems collect data on environmental conditions, equipment performance, and worker behaviors. This data is analyzed by AI to detect potential hazards such as overheating machinery, gas leaks, or unsafe worker proximity to dangerous zones. For example, sensors can alert operators when vibration levels in equipment exceed safe thresholds, preventing failures and injuries. Simultaneously, wearables can monitor worker fatigue or stress levels, issuing alerts to supervisors before accidents occur. By integrating these technologies, situational intelligence ensures dynamic risk assessment and immediate hazard detection.

AI Related Products and Services that could be Offered to Advance Manufacturing

To leverage the transformative potential of AI in improving both the physical and economic environments of manufacturing organizations, a variety of AI-related products and services can be developed. These offerings would focus on enhancing safety, efficiency, productivity, and overall business sustainability. Here’s a breakdown of key AI-driven solutions:

AI-Driven Products

1. Smart Sensor Systems:
   • IoT-enabled devices with embedded AI for real-time monitoring of environmental conditions (temperature, humidity, noise, etc.) and equipment health.
   • Wearables for employees that track vital signs, motion, and exposure to hazardous substances, providing personalized safety feedback.
2. Autonomous Robotic Systems:
   • AI-powered robots and cobots for high-risk tasks, such as material handling, assembly, or welding in extreme environments.
   • Drones for facility inspections, identifying structural issues, or monitoring safety compliance in large-scale operations.
3. Predictive Maintenance Platforms:
   • AI software for analyzing machinery performance data to forecast potential breakdowns and schedule maintenance proactively, reducing downtime and repair costs.
4. Digital Twins:
   • Virtual AI-driven replicas of manufacturing facilities for simulating and optimizing workflows, testing safety measures, and identifying inefficiencies before physical implementation.
5. Adaptive Safety Systems:
   • Dynamic AI tools that adjust safety protocols in real-time based on changing conditions, such as traffic patterns in warehouses or machinery reconfigurations.
6. Augmented and Virtual Reality Tools:
   • AR/VR solutions powered by AI for immersive safety training, scenario planning, and remote troubleshooting.
7. AI-Powered Environmental Controls:
   • Systems that optimize energy use, reduce emissions, and manage waste, leveraging AI to balance sustainability goals with operational efficiency.

AI-Driven Services

1. Consultative Safety and Efficiency Audits:
   • AI-powered analytics services to assess current manufacturing practices, identify risks, and provide actionable recommendations for improving safety and productivity.
2. Predictive Analytics as a Service:
   • Subscription-based platforms that analyze operational data to predict demand fluctuations, inventory needs, and production schedules, improving economic resilience.
3. Custom AI Model Development:
   • Tailored AI solutions designed for specific manufacturing processes, such as special hazards detection and reporting, unique quality assurance systems or bespoke automation tools.
4. Training and Workforce Augmentation Programs:
   • Services offering AI-powered adaptive learning platforms to upskill workers in using advanced manufacturing technologies safely and effectively.
5. Real-Time Incident Response Systems:
   • AI-enabled 24/7 monitoring and response services to detect and address safety incidents or operational disruptions immediately.
6. Regulatory Compliance Support:
   • AI solutions that track and analyze compliance with safety, environmental, and industry regulations, offering recommendations for maintaining standards and avoiding penalties.
7. Sustainability Optimization Services:
   • Consulting and AI tools to help organizations implement energy-efficient practices, reduce waste, and meet sustainability certifications, enhancing their economic and environmental standing.
8. Industrial Hygiene Intelligent Management System:
   • Utilization of a cloud system that provides deep data analysis of IH data, develops sampling strategies based on each specific work situation being studied, and conducts real-time monitoring with instantaneous analysis and communication of results and alerts.

Integrated AI Platforms for Holistic Management

Organizations could benefit from all-in-one platforms that integrate various AI functionalities into a centralized system, such as:

• Safety Command Centers: Real-time dashboards combining sensor data, predictive analytics, and incident response systems.
• Operational Digital Twins: Unified virtual environments that simulate safety, logistics, and productivity simultaneously.
• AI-Driven ERP Extensions: Enhancing enterprise resource planning systems with AI for better decision-making, from procurement to delivery.

By offering these products and services, AI providers can help manufacturing organizations create safer workplaces, optimize operations, and achieve economic and environmental sustainability, fostering innovation and long-term growth

I had a fascinating conversation building on this topic of products and services for EHS in Industry with ChatGPT. Some of this output is given above but I pushed further to some interesting insights. See the full chat at: https://chatgpt.com/share/67454451-e6cc-800a-8554-60399b176b0e

Ethical Considerations and Privacy

As with any technology, the use of AI in safety management raises important ethical considerations. One of the primary concerns is data privacy. AI systems often rely on large amounts of data to function effectively, and this data may include sensitive information about employees and operations. It is crucial to develop processes that protect individual privacy while still leveraging the benefits of AI.

Utilizing AI in a manner that preserves employees personal privacy and dignity is now a crucial area of development on the path to recognizing the full potential of AI to improve the workplace. The collection and use of employee data by employers must adhere to a clear code of ethical standards that all agree with. This is needed to reduce employee resistance and ensure anticipated regulatory compliance. Employers must implement an Employee Data Bill of Rights that covers data collection and utilization subjects such as: 

• Purpose and Benefit – The business purpose for collecting the data.
• Minimization – Only collecting what is needed for legitimate purposes.
• Fairness – No groups are excessively or arbitrarily targeted.
• Awareness – Employees are aware when, where and why data is being collected and utilized.

Harvard Business Review has a great piece on the boundaries between personal and employee data: https://hbr.org/2023/04/why-your-organization-needs-a-bill-of-rights-for-employee-data

Another ethical consideration is the potential for bias in AI algorithms. If the data used to train AI models is biased, the insights generated by these models may also be biased. This can lead to unfair treatment of employees or incorrect assessments of risk. To address this issue, it is essential to use diverse and representative datasets and to continuously monitor and adjust AI algorithms to ensure fairness and accuracy.

Primed by the above discussion on ethics and privacy, I collaborated with ChatGPT to develop several versions of a standard for the topic of Ethical Use of Artificial Intelligence (AI) in the Workplace: https://chatgpt.com/share/6745ba7d-001c-800a-89f7-4a01a2ecf816

Establishing Standards for AI in Occupational Safety

A key area of needed development in the application of AI in occupational safety is the establishment of standards for its use. These standards would provide guidelines on how to effectively implement AI in safety management processes, ensuring that the technology is used ethically and efficiently. This includes defining best practices for data collection and analysis, establishing protocols for monitoring and evaluating AI performance, and developing training programs to help employees understand and use AI tools effectively.

The development of a comprehensive consensus standard for the use of AI-enabled systems in occupational safety and health is an urgent and vital next step. In today’s rapidly evolving technological landscape, the integration of AI in workplace safety protocols is not just beneficial but necessary to ensure the well-being of employees. However, AI applications and methods applied to OSH is still developing, the best practices and industry consensus is yet to develop. Until these use examples have been fully developed and stabilized, the development of a consensus standard will continue to be a work in progress.

I dedicated significant time and effort to identify the key elements that such a standard must encompass. With the assistance of ChatGPT, I meticulously crafted a draft that addresses the complexities and nuances of this critical issue.

Following several rounds of revisions and expansions, I have arrived at a concept for a highly effective standard. This standard encompasses core principles such as real-time hazard detection, predictive analytics for potential safety risks, continuous monitoring and feedback loops, and ethical considerations for the deployment and use of AI technologies. By establishing such a comprehensive framework, we can enhance workplace safety, reduce accidents and injuries, and ultimately foster a healthier and more secure work environment.

The journey to develop this standard has been enlightening and challenging, but the result is a robust and adaptable blueprint that can guide organizations in leveraging AI for occupational safety and health. I am excited to share this concept and look forward to further discussions and collaborations to refine and implement these guidelines in practice. You can access the draft at this link: ISO Standard for AI-Enabled Systems in Occupational Safety and Health (OSH)

Unique Perspectives and Contributions

My extensive practice in the field of occupational safety and my five-year collaboration with Benchmark Gensuite have given me a unique perspective on the development and implementation of AI in safety management. This experience has allowed me to see firsthand how AI can transform safety practices, and I am committed to sharing these insights with the broader EHS community.

In addition to my work with Benchmark Gensuite, I have actively engaged with the EHS community through presentations and discussions on the impact of AI on safety management. My involvement with the ASSP Executive Safety Forum has provided a platform to share best practices and innovations, helping to drive the adoption of AI in the safety profession.

Staying Informed and Adaptive

To stay current with AI innovations and developments, I regularly collaborate with the AI team at Benchmark Gensuite to enhance and create new AI methods. I also follow industry leaders, engage in discussions, and read extensively on the topic. This continuous learning and adaptation ensure that I am always at the forefront of AI advancements, ready to apply these innovations in practical ways.

I will share a great source of the latest tech news and innovations that I found. It’s the TLRD AI newsletter: https://tldr.tech/

Passion for Progress

I am passionate about the ongoing adaptation of our profession to the needs of tomorrow’s workforce. The rapid pace of technological advancement means that the safety profession must continuously evolve to stay relevant and effective. By leveraging AI, we can empower EHS professionals to make better decisions, improve safety outcomes, and drive substantial improvements in workplace health and safety.

One of the key outputs I hope to achieve through my involvement with the AI Task Force (if selected) is the delivery of actionable insights and recommendations for occupational safety organizations, such as ASSP, and industrial employers. These insights will help EHS practitioners leverage AI technologies to enhance their professional mission and positively influence the future of work.

The Mercurial Nature of AI Clouds Understanding of Full Potential

AI’s mercurial nature—its rapid evolution, unpredictable advancements, and shifting capabilities—poses significant challenges for establishing stable, real-world applications. The technology’s continuous and often unexpected growth makes it difficult for regulatory bodies and industry leaders to create long-term policies. By the time a framework is established, AI may have already outgrown it, leading to regulatory gaps in critical sectors like finance, healthcare, and industrial safety. Additionally, AI models trained in controlled environments frequently struggle with real-world edge cases. Factors such as changing environmental conditions, unforeseen safety risks, and complex human interactions can result in unreliable performance, which is particularly concerning for safety professionals integrating AI into industrial risk management.

Beyond these immediate challenges, AI’s unpredictable trajectory also makes it difficult to fully grasp its long-term potential. Because its capabilities are constantly evolving, industries struggle to envision all the ways AI could transform their operations. Many organizations implement AI only for narrow, well-defined tasks, missing opportunities for broader, more integrated applications that could drive efficiency and safety. This uncertainty also contributes to hesitation in investment and adoption, as businesses are unsure whether the AI tools they deploy today will remain relevant or be quickly surpassed by new advancements.

Ethical and legal concerns further complicate AI adoption, as issues like bias, accountability, and transparency remain unresolved. In safety and sustainability, for example, AI-driven decision-making can conflict with regulatory compliance if it lacks explainability. At the same time, the rapid pace of AI development outstrips workforce training efforts, creating skills gaps and resistance to adoption. Industrial workers may hesitate to trust AI-driven safety monitoring tools if they don’t understand how they function or if past implementations have been inconsistent. Compounding these issues, AI is highly dependent on data quality, but real-world applications often struggle with incomplete, biased, or dynamic data sources. An AI model trained on one facility’s incident reports may not generalize well to another with different equipment and processes, leading to unreliable performance.

To navigate these challenges, industries are prioritizing hybrid AI-human collaboration, ensuring AI augments rather than replaces human decision-making. They are also working on adaptive regulatory frameworks that can evolve alongside AI’s capabilities, developing robust validation processes to test AI in real-world conditions before full deployment, and emphasizing explainability and transparency to build trust in AI-driven applications. However, the fundamental challenge remains: AI’s unpredictable evolution makes it difficult to see its full possibilities, leaving businesses and policymakers struggling to balance short-term implementation with long-term vision.

Combating AI System Over-Reliance and Skills Atrophy

Applying Human-in-the-Loop (HITL) and Human-Centric Design (HCD) in AI-driven Occupational Safety and Health (OSH) systems ensures that automation enhances, rather than replaces, human expertise, leading to safer and more effective workplaces. HITL keeps safety professionals actively engaged by requiring human oversight in critical decision-making, reducing automation bias and ensuring that AI-generated insights are validated before implementation. This approach enhances situational awareness, prevents over-reliance on automation, and allows for adaptive responses in dynamic or high-risk environments. By integrating human expertise with AI’s analytical capabilities, OSH professionals can proactively identify hazards, assess risks, and implement controls that are both data-driven and context-aware.

Human-Centric Design (HCD) ensures that AI systems in OSH are intuitive, transparent, and aligned with worker needs, increasing user trust and adoption. By designing AI interfaces that present clear, actionable insights rather than overwhelming users with complex data, HCD enhances decision-making efficiency and effectiveness. Additionally, AI-driven safety systems designed with HCD principles can improve training, emergency response, and ergonomic risk assessments by adapting to user behavior and providing real-time feedback. This leads to more resilient and adaptive workplaces, where human workers and AI systems collaborate seamlessly to mitigate risks, reduce incidents, and promote a culture of continuous safety improvement.

There is a rapidly growing body of knowledge and research in this area. See the conversation I had with ChatGPT on this subject: Human in the Loop (HITL) Applied to OSH Applications

Conclusion

The integration of AI into Safety and EHS management holds immense potential for transforming how organizations operate. By improving decision-making processes, enhancing safety practices, and ensuring ethical considerations, AI can drive significant improvements in workplace safety and health. My collaboration with Benchmark Gensuite and my engagement with the EHS community have given me valuable insights into the practical applications of AI, and I am committed to sharing these insights to help advance the safety profession.

By staying informed about AI innovations, promoting ethical standards, and fostering a culture of continuous improvement, we can ensure that AI becomes an essential tool for the EHS profession. Together, we can leverage AI to create safer, more efficient workplaces and drive the future of occupational safety and health.

Microsoft Copilot’s summary of this article

The expanded article provides an in-depth look at the transformative potential of AI in EHS management, drawing on practical experiences and collaborations to highlight the importance of leveraging AI for decision-making, safety improvements, and ethical considerations. It emphasizes the need for standards and continuous learning to ensure the effective and responsible use of AI in the safety profession.