Three developments shaping how safety is built in modern industry right now
Safety in industry is becoming more clearly defined, not primarily through new principles, but through clearer expectations of how technology should function in practice. This is particularly evident in how robot standards are being applied, how the EU is prioritizing established technologies, and how AI is becoming a more integrated part of safety architectures. For Cedoc users, this largely comes down to understanding how systems work together and how this should be described in risk assessments and documentation.
Robot standards in practice
How prepared are today’s robotic systems for the requirements in ISO 10218:2025? That question is explored in a recent industry analysis of how organizations are working to meet the standard in real-world applications.
The article highlights that much of the required technology is already in place, but that the focus is shifting toward how safety functions in operation, especially in environments where humans and robots work closely together. Kinematic monitoring is highlighted as a key element, meaning the continuous tracking and control of a robot’s movement, speed, and position.
What this means in practice:
- Safety needs to be described in terms of how systems behave in real conditions
- Risk assessments need to capture the interaction between robot, control system, and work tasks
- Functions need to demonstrate how they actually reduce risk, not only that they are in place
For Cedoc users, this points, among other things, to the importance of connecting risk, mitigation measures, and actual system behavior more clearly.
Learn more about the standard and access the two parts here:
ISO 10218-1:2025, Part 1 – Industrial robots
ISO 10218-2:2025, Part 2 – Industrial robot applications and robot cells
Photonics takes on a larger role in industry
The EU’s focus on photonics does not primarily introduce a new technology area, but rather reflects how an established technology is taking on a broader and more integrated role.
Photonics, where light is used to measure, analyze, and control, is already widely used in many applications. What is changing is how the technology is increasingly integrated into larger industrial initiatives.
What this means:
- Systems are becoming more sensor-driven and increasingly reliant on data
- Functions are being connected into larger technical systems
- Precision increases, while systems also become more complex
What this means for Cedoc‑users:
- Risks more often arise at the interfaces between systems
- There is a greater need to understand how data influences function and decision-making
- Documentation needs to show how system components are connected in practice
Safety systems for physical AI
With solutions such as NVIDIA Halos for Robotics, it becomes clearer that safety is no longer something separate, but part of the overall system design.
The article describes how AI, sensors, and safety functions are integrated into a unified structure, where decisions are made in real time. These types of solutions are already being used in industrial environments today.
What is becoming clearer:
- Safety systems are becoming more adaptive and data-driven
- Decisions are made based on sensors and AI models in real time
- Function and safety are combined into a unified system logic
For Cedoc users, this means:
- Risk assessments need to describe how systems act, not only what components are present
- Focus shifts to how entire systems interact
Documentation needs to capture dependencies and interactions
What is worth following
- How organizations continue to work with the application of ISO 10218:2025
- How photonics becomes more integrated across different industrial systems
- How practices evolve around AI-based safety in physical environments
In summary, the main development right now is less about new requirements and more about making existing requirements work in practice, within increasingly complex systems. For Cedoc users, this points to a growing focus on describing how systems function in reality, from technical setup to real-world behavior.
