Using AI in EHS: AI Use Cases and the Future of Intelligent Tech with Cority

AI use cases webinar recap blog starry night photo cority

As artificial intelligence (AI) rapidly reshapes industries, questions about its practical applications, challenges, and future potential are more prevalent than ever.  

From enhancing workplace safety and efficiency to ethical considerations and strategic innovations, AI is transforming industries and it’s crucial for businesses to understand where it’s headed next. In this blog, we will aim to answer the following questions on the future of intelligent technology:

  1. What is AI?
  2. How are organizations approaching the question of AI in EHS?
  3. What do users look for from AI?
  4. What are some common use cases for AI within EHS?
  5. What is the biggest obstacle organizations face when looking to bring AI into their systems?
  6. For organizations looking to enter the AI space, what is their first step?

What is AI?

AI can be broken down into three broad categories, two of which have been around for a while, analytical AI and predictive AI. The third, being Generative AI, is what has caught so much attention recently.

Analytical AI

Analytical AI is often confused with data analytics, which is the ability to analyze datasets and provide specific insight. For example, if there is a list of addresses in a dataset, analytical AI can hypothetically be used to create a subset of data for all addresses that are within a 50-mile radius of a given point. Many of us use analytical AI frequently. 

Predictive AI

The next category is predictive AI, where based upon trends in a dataset, an AI model can be used to estimate what future values of that dataset would be. It can create values based upon the dataset or use probability to estimate future values or ranges. Let’s say someone wanted to track fitness – in the first week of tracking, they run 3 miles, in week two they run 4 miles, and in week three they run 5 miles. In this instance, predictive AI could be used to estimate how far they would run in the future based upon that dataset. 

Generative AI

Finally, the AI category that is capturing so much attention now is generative AI, which is AI that uses large language models (LLMs), trained on very large datasets to detect and track patterns within that dataset, and use those patterns to generate new content. LLMs can also be of images and other media. Generative AI can be trained on millions of images of cats and text that people have used to define each picture and then use that data to create a new image and description of a cat when prompted by a user. 

With regard to Generative AI, this technology is still in its infancy. Cority’s interpretation and goal for any AI research and development is to create technologies that augment human capabilities and improve overall productivity, efficiency, and quality of life. In other words, when Cority thinks about computer systems and algorithms that utilize AI, they believe software providers developing AI tools should be focused on those that enhance and extend human abilities.  

AI technologies are designed to complement human skills by automating routine tasks, providing insights from large amounts of data, assisting in decision-making processes, and enabling new forms of interaction and communication.  

How are organizations approaching AI in EHS?

Generally, both organizations and individual users approach conversations about AI with optimism, but also caution. When it comes to expectations, there can be a dichotomy between “it will change everything” and “there’s risk to consider”.   

Typically, organizations are currently looking to automate highly manual and repeatable tasks specifically around data collection, quality assurance and control, notification of escalation workflows, or aggregation and analysis type tasks. 

What do users look for from AI?

Users seek actionable insights from AI-driven data analyses that incorporate past performance and trends. Trending and insights will be primarily driven from internal data sources and can be expanded to include external data sources in analyses. However, it’s important to remember that this technology is in its infancy; the things users want to do might be far more complicated than they think. Additionally, overall trust in AI needs to grow. Use cases need to start simple so that users gain confidence in AI Tools, and as use grows, the complexity of the tool and insights being delivered can and will grow, too. Clear value at the individual level needs to be demonstrated in order for adoption to occur. 

How do goals and expectations vary from front-line users?

Typically, organizations are seeking measurable value and individuals are looking for use cases that create personal benefits. There is a need for co-creation when exploring the use of AI and, put simply, you can’t have one without the other. Organizations must provide the framework for responsible AI use as well as the resources to enable it, but individuals will be the ones using AI and creating value.   

Organizations must understand the use cases most relevant to their employees and focus on enabling them. Additionally, organizations must understand the level of maturity of their people when it comes to AI and how readily it will be adopted, meaning education and training will be required.

What are some common use cases for AI within EHS?

Recent research published by Verdantix has shown that there are 5 distinct use cases within EHS/ESG that are most desired by users;  

  1. AI Assistants: Providing data analysis and insights or performing summary generation  
  2. Computer Vision: Identifying unsafe behaviors and non-conformances, as well as 3D motion capture (often within Ergonomic spaces) 
  3. SDS Automation: Automating Safety Data Sheet indexing 
  4. Permit Deconstruction: Extracting information from free text permits of varying formats, and; 
  5. SIF Prediction: Identifying events and allocating resources for reductions in said events 

What is the biggest obstacle organizations face when looking to bring AI into their systems?

One of the biggest issues for organizations to consider is the quality or the quantity of the data in their systems today. As with many things in software, you get out what you put in, and that very much applies to raw data. Systems that utilize AI can do some pretty amazing things, but in any capacity, it relies on data to be accurate to do so. 

For Generative AI, there are plenty of additional points for consideration. Many organizations either don’t want to or can’t use public learning models. This can be for any number of reasons, and will likely change over time but, for now, organizations want solutions that learn quickly or are context aware straight away. In many cases, this simply isn’t possible without the use of LLMs that are based on public learning models.  

Finally, there’s a common expectation for technology to be right every single time. We get mad when our computers freeze, or when our GPS calls out an exit on the highway too late – and the truth is that this is an unfair expectation to place upon new technology. There is significant time and effort that goes into teaching AI models to complete specific functions – and to ensure AI isn’t teaching itself based on incorrect assumptions or false data. As such, there is absolutely a level of patience required for organizations looking to bring Generative AI into their systems. 

For organizations looking to enter the AI space, what is their first step?

As with any software implementation, the first step is to identify the problem statement. If organizations can ensure that everyone on the team is on the same page when it comes to goals, they stand a far better chance of realizing that goal. Ask what it is your organization wants AI to help with There are likely dozens of questions further down the road, including how to implement the solution, identifying a partner to help achieve this vision – but the end goal must be fleshed out or it’s not going to be achieved. 

Once there is clarity around the use cases, it is critical to enable the framework and governance for the responsible use of AI and develop the data architecture that will accelerate the ability to implement and scale AI. 

In Cority’s on-demand webinar, Rethinking AI: Dive into the Future of Intelligent Tech, they discuss the EHS market’s current perception of AI, insight into how the topic is being broached by both organizational leaders and end-users, and provide examples of how technology is currently being leveraged in existing EHS & ESG ecosystems.    

To check out the full rundown of how organizations can make the most of the tools and techniques available with AI, watch the on-demand webinar today! 

Recent/Latest Blog