The future of industry: AI becomes a product in its own right

"Industrial companies must consistently view their products from two sides: On the one hand, the physical machine or system, and on the other, the associated AI as a practically independent product to be integrated," says Jens Rollenmüller, Regional Vice President at Aras, a provider specializing in Product Lifecycle Management (PLM). He believes that AI must be considered on an equal footing with mechanical, hydraulic and electronic components.

Industrial machines and systems are equipped with numerous sensors that constantly record operating data. In the future, integrated AI will evaluate this data in real time and actively intervene in the operating process if necessary.

Why did the AI decide when and how?

Jens Rollenmüller, Regional Vice President at Aras. © Aras Software

Artificial intelligence is therefore increasingly automating decision-making processes and taking on thinking tasks that were previously reserved for humans. According to industry expert Rollenmüller, this development triggers an obligation to provide evidence: "Essentially, it's about the question of why the AI decided when and how. Because when machines act autonomously, it is the responsibility of companies and manufacturers to analyze and document these decision-making processes - and thus possible sources of error - completely for all components and products used."

To ensure this transparency, AI should not be viewed as a feature of a machine, but as an independent product or component. "This approach speeds up development cycles while at the same time ensuring structured data management and compliance with all regulatory requirements, such as the EU's AI Act," says Rollenmüller. "The development or further development of AI is far more complex than the software components of traditional automation technology with their limited functionalities. AI must function autonomously, safely and error-free in many different environments and scenarios."

The higher the degree of autonomy of the AI, the more important cross-departmental collaboration becomes - from research and development to production and assembly through to IT and the compliance department. This requires maximum transparency and traceability, efficient data management and clearly defined responsibilities. "Modern and, above all, flexible PLM systems have already shown in the past how increasingly complex products and production environments can remain manageable. This will also be the case with AI," Aras manager Rollenmüller is convinced.