Should safety professionals be cautiously optimistic about using AI?

The benefits and challenges of leveraging artificial intelligence

Should safety professionals be cautiously optimistic about using AI?
Abdel-Rahman Sabriye

One of the worst-kept secrets has been the massive rise in awareness and usage of Artificial Intelligence (AI), which has recently dominated headlines. AI is the computer science field that develops and studies intelligent machines that mimic human intelligence.

There has generally been massive excitement with the development of AI and its potential to transform the world, much as the introduction of the internet did back in the '90s. 2023 saw the most rapid uptake in AI technology, with applications such as ChatGPT and LaMDA now becoming household names. Interestingly, despite the recent popularity, AI has been around for a while, with popular applications such as Siri and Alexa. 

AI is expected to spread to most industries to help make everyday tasks quicker, smarter, and more efficient. However, with this excitement, an understandable level of hysteria followed. Hysteria about technological advancements is not new, especially when we do not understand the full implications. Fears of technological singularity aside, it is essential to question whether and how AI can revolutionize the world of safety for safety professionals.

I will delve into the anticipated benefits AI will try to achieve and the challenges that will undoubtedly be associated with this undertaking.


1. Improve Hazard Detection for High-Risk Industries

There is hope that generative AIs can use advanced algorithms to identify potential safety hazards through pattern identification before they become dangerous safety issues. This application is already in use, such as cameras or drones with AI applications. This can greatly minimize the personal safety risk to individuals and ensure the detection of potential safety issues, especially in high-risk industries. 

2. Predictive Analytics

With all the information that AI systems require to be successful, AI systems can also identify patterns and trends that may be useful in indicating potential hazards within the workplace. This can be used proactively to develop new safety measures and better adjust existing ones to prevent accidents.

3. Supplement Safety Documentation 

To the chagrin of many safety professionals, safety is rife with documents, policies, SOPs and other documentation. This often can take a significant amount of time and, therefore would be some utility to reclaiming this time. The article “Should we let AI author our safety documents?” looked into this and determined although using AI to create safety policy documentation had several flaws, it did provide a good starting point for SMEs to review and ensure it is relevant to the industry and organization.

4. Learning and Training

Training plays a big part in OHS to ensure workers understand how to identify hazards, protect themselves and react accordingly. AI’s can potentially revolutionize training through different avenues. AI-generated content creation through AI-generated videos, which was recently released such as OpenAI Sora is a tool that uses AI to create videos f

rom written prompts. This can potentially revolutionize how training is provided. There is also personalization of learning based on individual learning needs. AI can also help identify learning gaps through AI-powered data analysis and metrics, which create profiles based on these gaps to ensure that they are filled with appropriate training.

With all this in mind, the future of AI integration within safety is encouraging. However, there will also be several challenges that AI must overcome before being introduced into workplaces in the near future.


1. Ethical Implications  

The ethical implications of AI will be one of the significant issues to contend with. This can range from several issues but typically surrounds problems with data privacy and sensitivities. Since AI often requires vast amounts of information, this can include personal information that must be protected. AI can require personal monitoring, which can have a perceived negative effect on employee wellbeing. There may be concerns regarding data bias/discrimination. Organizations will have to put in significant policies to ensure no misuse, abuse or potential breaches. This means robust and costly cybersecurity measures to prevent this.

2. Data Availability

AI typically requires extensive data sets to ensure accuracy and limit biases. Without this crucial step of providing accurate and carefully selected information, the system may not be set up for success. The results you get from the system are only as good as the data you put in and are available. Most organizations may not have this readily available, so they cannot realize the full potential of AI.

3. Threat of Replacement

The perceived threat of technology replacement among employees has been shared with most technological advancements. This narrative is likely to continue with the rise of AI; however, AI will likely change our process, and employees will need to harness its power. Communicating this will be extremely important to businesses to ensure employees know it is there to enhance and not replace. 

What’s Next?

In conclusion, although there will be several major challenges with implementing AI within the safety field, it is promising, and safety professionals should be cautiously optimistic. AI can never replace the human element of OHS, but it can undoubtedly enhance the field, empowering safety professionals with a powerful, productive tool to better serve their organizations.