On October 16, 2024, we once again focused on artificial intelligence. This time, the focus was on the opportunities arising from its use in the field of risk assessment.
To address the topic of AI and risk assessment, 12 managers from different functional areas came together in our port room to take an in-depth look at innovative approaches.
Martin Westhoven and Arn Baudzus from the Federal Institute for Occupational Safety and Health prepared and led this workshop for us.
After a round of introductions of all participants, there were two theoretical keynote speeches by our speakers, which referred to current research projects. First, Martin Westhoven discussed how Natural Language Processing (NLP) can be used to improve risk assessment. There is significant potential here for an AI application if text data is primarily available and speech is to be processed to support risk assessment.
The second presentation by Arn Baudzus dealt with how information from environmental parameters (noise, gas concentrations in the air, temperature, etc.) can be used to support risk assessment. Data on temperature, air pressure, particulate pollution, gas concentrations etc. can be collected via a sensor network. Anomalies in the determined environmental parameters then provide indications of a risk assessment that needs to be improved. The second part of the workshop focused on specific application examples: Which typical processes in a risk assessment could be supported by AI? Or: What happens rather rarely, but has an even greater impact? In small groups, the participants played through different cases as scenarios to find out where the use of AI could be useful. It was particularly important to consider all stakeholders such as line managers, the respective team, IT administration, safety experts, etc. in the selected case study and to examine the individual process steps in detail. Finally, the small groups presented their case studies. The scenarios were discussed in plenary to find out how an AI system could help with the individual risk assessment.
Requirements are to be derived from this so that the development and testing of suitable prototype AI systems can be tackled in the future. NIRO would like to thank Martin Westhoven and Arn Baudzus from the Federal Institute for Occupational Safety and Health for this interesting and instructive workshop, which provided new insights into the potential of AI applications in the field of risk assessment. We will continue to work on this topic – and are excited to see what research will find out to support it.
Martin Westhoven and Arn Baudzus (from left to right) gave the group a very competent and understandable introduction to the innovative approaches to AI in the field of risk assessment. Picture: NIRO e. V.
