AI accelerates pace on the factory floor
What does generative AI on the factory floor refer to?
Generative AI is a technology that with the help of deep neural networks develops creative solutions by recognizing complex patterns in data and independently generates creative new designs, processes, or products, or optimizes existing ones. That’s not just about basic automation of processes that can be achieved by other AI methods as well but about actively helping to shape and enhance the entire manufacturing landscape. AI becomes a “colleague,” a “co-designer.”
Generative AI can achieve major optimizations especially in complex, multi-level manufacturing processes, primarily where many variables must be controlled at the same time. AI can dynamically adjust production parameters in real time that previously were statically posted or had to be manually adjusted by highly experienced skilled workers to minimize scrap and to maximize product quality. These technologies have already achieved impressive results in many industries.
How do the capabilities of AI exceed those of classic computer programs particularly on the factory floor?
Unlike classic computer programs, AI, through continuous learning and adapting from large data volumes, can make more precise predictions and respond flexibly to unexpected situations. This capacity to be based not only on rules but on experience makes AI particularly valuable in dynamic manufacturing environments.
An outstanding example is the utilization of generative AI for topology optimization in additive manufacturing such as 3D printing in which components are structurally improved to reduce weight while strength increases at the same time.
An example from the automotive industry is the optimization of production lines in vehicle assembly settings. Here, AI can be used to adjust material flow, cycle times, and the sequence of work steps in ways that accelerate production speed while reducing energy consumption. Another example from the automotive supplier industry is the optimization of injection molding methods for plastic components. Generative AI can adjust parameters such as injection pressure and cool-down times in real time to minimize defects in shape and to maximize the quality of the parts produced. Such optimizations lead to lower scrap rates, shorter production cycles, and generally more efficient manufacturing.
How specifically does the integration of generative AI affect the competitiveness of companies? And does AI know-how on the factory floor come from internal or external sources?
The integration of generative AI can considerably enhance competitiveness by optimizing and flexibilizing production workflow, and by accelerating innovation processes. Due to automation and dynamic adjustment of production parameters, setup times are shortened while product quality is kept at constantly high levels, which leads to higher production volumes. These efficiency gains enable companies to reduce costs and to offer competitive prices, which provides them with a decisive advantage in the global marketplace. However, it must be noted that these methods are basically available to all companies – and that even companies with less previous knowledge or fewer well-trained employees can achieve very good results with them much faster. The existence of these methods potentially increases competition.
AI know-how on the factory floor often emerges from a mix of internal and external knowledge. Companies that already invest intensively in research and development establish in-house competence centers while others rely on external experts and specialized AI service providers. Collaboration with universities and research institutes plays a crucial role as well when it comes to staying at the top of one’s game in terms of technological developments.
Are manufacturing processes and jobs going to fundamentally change due to the advance of AI?
Yes, especially due to cobots that – assisted by generative AI – can flexibly adapt to a variety of tasks and interact with people in a symbiotic work environment. That leads to higher productivity, better ergonomic conditions at the workplace, and reduction of errors because AI-controlled cobots can learn and achieve optimizations in real time.
The rapid development of AI confronts the ways in which we’re going to interact within teams in the future with a fundamental transformation. AI is not only a tool for automating many processes but an essential driver of new collaboration models far exceeding what we’ve known so far. Hybrid teams combining human and artificial intelligence offer an exceptional opportunity to redefine the boundaries of what we can achieve together.
“The advance of AI is going to fundamentally change manufacturing processes, especially due to the increased deployment of collaborative robots working closely together with human workers.”
Prof. Dr. Sabina Jeschke
An AI error brings production to a halt – who’s going to be liable?
The question of liability in the case of AI-based systems is currently – in ALL areas, not just on the factory floor – a complex and dynamic field encompassing both technological and legal aspects.
Basically, liability in the event of an AI error depends on where the error occurred and who has control of the system. If, for instance, the error can be attributed to improper use or a wrong configuration of the AI system by the company the company as the operator could be held liable. However, if the error is to be attributed to a defect in the AI software or a malfunction for which the manufacturer of the system is responsible liability would pass to that manufacturer, like in the case of a vehicle with a technical defect.
In practice, liability is often subject to contractual arrangements defining in detail who will be responsible in what cases. Also, in some jurisdictions, there are discussions about new laws to be developed specifically for AI systems so as to create clear liability rules. It’s highly probable that the statutory frameworks are going to see an evolution in coming years to adequately deal with the challenges posed by AI.
What regulatory hurdles exist worldwide concerning the use of generative AI on the factory floor?
There are several regulatory hurdles worldwide for the use of generative AI on the factory floor that companies must observe. In the area of data protection, companies must ensure that all data collected and processed for training the systems comply with applicable data protection laws such as the General Data Protection Regulation in the EU and that they’re protected against unauthorized access. In terms of product liability, the question arises about who will be liable if defective products are made due to an AI decision, which calls for clear statutory frameworks. Product safety standards require products that have been co-designed by AI to meet all statutory safety requirements to protect consumers. Moreover, the safety standards of the production process, including industrial health and safety, must be looked at, requiring AI systems to be designed in ways that exclude risks to employees. In reference to intellectual property and copyright there are challenges when AI creates designs of its own because the legal assignment of property rights is not always clear. Finally, industry-specific regulations must be observed as well that may impose additional requirements on the utilization of AI, depending on the industry such as automotive or pharma.
What costs does the implementation of generative AI on the factory floor entail?
The implementation of AI on the factory floor entails considerable capital expenditures that will pay off in the long run, though. They must be identified on a case-by-case basis; there’s no “rule of thumb.” Costs include purchasing the necessary hardware, development or customization of software, integration in existing systems, and employee training. In addition, costs of ownership are incurred for maintenance, updates, and data management. Despite these initial investments, companies can achieve significant savings and competitive advantages because of higher efficiency, reduction of defect rates, and better product quality.
Talking about sustainability: How can AI on the factory floor have a positive effect on nature?
The positive effects of AI on nature are not limited to those on the factory floor. AI systems in general can make positive contributions to sustainability by optimizing processes that use resources more efficiently. For instance, AI systems can minimize energy consumption by precisely controlling machines and adjusting production parameters in real time. In addition, AI can help reduce waste by optimizing material consumption and reducing scrap. Due to actions like these, manufacturing industries can reduce not only their environmental footprints but also lower the pressure on natural resources which, in the long run, leads to more sustainable manufacturing.
How long is it going to take for AI to evolve from apprenticeship to mastery?
It’s true that – despite the impressively fast progress made especially in generative AI – the methods are still at an early stage. It’s difficult to name an exact time span but I expect AI to rise from apprenticeship to mastery in many manufacturing areas in the next three to five years. The technology is developing rapidly and with increasing maturity and integration into industrial processes, AI is going to become more and more autonomous and capable. However, full mastery is not going depend only on technological progress but also on how well humans and machines can work together and how fast companies are going to be willing to adopt these changes.
The apprenticeship-mastery question is particularly important because it illustrates that knowing what capabilities AI has today or does not (yet) have is not enough. For a company to have a sustainable overall AI strategy it’s crucial that top management can judge what AI will be able to do in the future and when.