Artificial intelligence has arrived almost unnoticed, from being present on the big screen to becoming part of our daily lives, and consequently of our work environments. Technological change is accelerating this transition with AI applications through cobots, assistance tablets, chatbots, and/or implementations in work protective clothing or individual protection equipment.
AI can create opportunities, but also new challenges for occupational safety and health (OSH), its management, and its regulation generate significant debate to which I would like to shed some light.
Differences between AI systems and traditional ones
We are facing a new scenario in Occupational Safety and Health. For example, Big Data allows us to use structured and unstructured data in the workplace to identify safety issues. External data can develop predictive analytic algorithms to identify risk factors that affect other companies and prepare us before they become a problem. This leads us to ask ourselves:
Can AI predict and manage occupational hazards better than traditional systems?
In my opinion, AI-based control can support prevention based on tested data but never replace traditional methods. The combination of both will have a significant boomerang effect. It will allow for more efficient and organizational-scale risk and OSH assessments and inspections. Trends can be analysed, details can be delved into, real values can be measured, and/or emissions performance can be monitored with just a few clicks, something that was impossible with paper records.
This is changing the way decisions are made within organizations, showing weaknesses and strengths. All this without forgetting that ethical and strategic decisions are needed with effective systems that handle a large amount of confidential personal data.
Ultimately, it is important to ensure transparency in the collection and use of such data. Additionally, staff and their representatives should be empowered through the same access to information and trained in its use and handling. Which brings us to the next point.
Training to receive about AI in the context of OSH
As data automation and analysis become more integrated into everyday life, there will continue to be a great need for problem-solving related to the application of this new perspective. Here we face three challenges in terms of training:
Safety professionals working collaboratively with the entire company using solid information from this data could improve their assistance by moving from predictive to prescriptive, thereby reducing exposure. This requires providing them with the necessary knowledge and skills for a beneficial digital transformation. And of course, this knowledge should not stop there, this analytical and continuous improvement spirit should be integrated into workers, middle managers, and top management.
Similarly, companies need to know the necessary information about these changes, their impact on the nature and organization of work, and the new challenges for OSH that they may entail. The lack of digital skills poses a risk of social and economic exclusion and therefore an impact on the mental health of the workforce. This gap may be motivated by various reasons: geographical location, economic conditions, level of education, age, gender, etc. In this scenario, clear policies and promotion of training actions focusing on the most vulnerable are needed.
And even more, the opportunity to feel the risk in a real way, the impact that experiencing a fall, drowning, asphyxiation, etc. would have on the worker should not be underestimated; clearly, the commitment to virtual training and emergency training should be adopted. But if it is not done, it is because OSH is seen in many organizations as a cost, which leads me to this reflection:
Reduction of costs associated with occupational accidents by AI
As an HSE manager in a technology company, it is common for prevention technicians to ask me for arguments to convince investment managers in OSH technology, especially in those organizations where OSH is not fully integrated. Because it is difficult to see the return on investment.
The cost-benefit measurement of non-prevention must be an intrinsic question in every business management plan. Most entrepreneurs have considered it as a regulatory obligation that had to be fulfilled, they do not see beyond. It is the task of OSH professionals to value preventive investment (improvement of absenteeism and incidents) and the productivity vs. well-being at work relationship.
Calculations by EU-OSHA (European Agency for Safety and Health at Work) showed that for every euro invested in promoting occupational health, almost five euros of benefit were generated. What better investment than anticipating risks and effects on occupational health by motivating the worker and giving him tools to progress in a new work scenario? Investing in Safety and Health is always a safe bet, towards this, we should direct the future.
Proper balance between AI and human supervision in occupational safety
To conclude, addressing a point that generates much debate today, what is the appropriate balance between automation through AI and human supervision to ensure safety in the workplace?
The use of AI to manage workers also poses numerous psychosocial risks: loss of control by workers over their jobs, increased intensity, and performance pressure, decreased managerial social support, individualization, dehumanization, and/or limited participation, all of which have a significant negative impact on mental health.
A recent report by EU-OSHA shows that it is necessary to have a strong "prevention through design" approach that integrates a human-centered perspective. A broader approach is needed in the design and use of these devices, it must be designed, implemented, and managed reliably, transparently, empowering, and understandable.
Safeguru | Delivering a Safer Future
Susana Romero Miguel, Environmental Health and Safety Manager - General Secretary at AEPSAL