At Datalyxt, we became involved with artificial intelligence (AI) applications at an early stage. It was more than helpful to have chosen Karlsruhe as the location for this. We took our first steps at the Forschungszentrum Informatik (FZI) in 2007. Over the course of many years of research and corporate projects, SdbHub emerged from this.
The evolution of SdbHub – first digitize the safety data sheet
Datalyxt develops information retrieval platforms powered by AI. Whether web, documents or emails, Datalyxt is able to read information from almost all accessible information carriers in a targeted manner and prepare it analytically.
One of our first use cases was a news analysis platform called knit (2014-2016, now in the care of another company). SonarBox followed in 2016, and we have been using it to supply companies with suitable information from the web ever since.
What we had to realize in the course of other projects: There are always requests for information on the safety data sheets, which were rather unknown to us until then. We quickly realized why the demand from suppliers is so huge. There are enough reasons for the digitalization of safety data sheets. Due to the lack of and standards, recipients have major problems processing safety data sheets quickly, with high quality and in compliance with the law. Without AI, this is not possible at all.
Occupational safety and safety data sheets
When working with AI, the first thing that comes to mind is neither occupational safety nor safety data sheets. The ball was set rolling in terms of occupational safety by a project in 2016. We were informed by the company AGU GmbH about a problem in the field of hazardous substance management. Within a few weeks, a prototype was ready. Of course, a lot has happened since then. It is one thing to receive a hint for a problem, quite another thing to successively design and build a product from it according to customer requirements.
Datalyxt was able to use the know-how it had accumulated over the years and apply it to safety data sheets (SDSs). This is how SdbHub was born.
Hub of SDSs
SdbHub consists of the word components SDB (Safety Data Sheet in German) and Hub. The SDB in SdbHub is not capitalized. There is a reason for this. Without SDB there is no basis for information, but without Hub there is no knowledge. Both elements are of equal importance and are interdependent in the information gathering process.
It is not just a matter of reproducing certain fields of an SDS. Nor is it just a matter of looking at safety data sheets. Rather, it is about a holistic view of occupational health and safety, in which SDS processing takes place. Data extraction is intended to provide users with analytical and fluent information within their processes and systems.
In the future, further solutions in the field of occupational health and safety will be developed under SdbHub and made available to companies and solution providers. The aim is to gradually make artificial intelligence methods accessible to occupational safety and health. In this way, we also help solution providers who are unable to muster their own resources for the development of AI solutions and thus strengthen the profile of their programs.
SdbHub sees itself as a back-end solution. We want to operate similarly to database management software. Therefore, it is very important for us to be used across platforms and vendors.
In the first step, we will provide SdbHub with both an initial quality check and a content extraction of a safety data sheet. The results will be provided in a structured way.
How SdbHub continues to grow
Our goal is to rethink occupational health and safety processes and develop new and more effective methods for doing so. These can be, for example, adaptive e-learning content or proactive assistants that recognize hazardous situations and inform employees about such hazardous situations.
