Artificial intelligence | SMEs need help with technical implementation

Artificial intelligence is still little used in production – especially in medium-sized companies. These challenges still exist in companies.

Artificial intelligence
Artificial intelligence

In industrial companies, the expectations of artificial intelligence (AI) are high, but it is still not used much in production – especially in medium-sized companies. There is ambiguity, for example, in the assessment of the economic benefit, the applicability in one’s own production environment as well as in the collection and use of data. This is the result of an online survey conducted by the AI ​​innovation project IIP-Ecosphere with the support of the project partner VDW.
The Federal Ministry of Economics is funding IIP-Ecosphere as part of the AI ​​innovation competition to accelerate the use of AI in production. The aim of the survey was to determine the current status and practical challenges of companies with regard to the use of AI and related topics such as data collection, cloud use, selection and framework conditions of AI solutions and Industry 4.0 platforms. “With the survey, we wanted to show necessary, real fields of action in which IIP-Ecosphere can now actively act as an accelerator for the utilization of AI methods,” explains Dr. Claudia Niederée from the L3S research center, project leader of IIP-Ecosphere. 75 companies took part in the survey, two thirds of them from the machine tool industry.
The podcast episode with AI expert Markus Ahorner is also about artificial intelligence:

More than half have not yet had time to deal with AI

What is striking is the high participation of larger and very large companies, at 70 percent. The background could be a greater preoccupation with AI topics compared to medium-sized companies. Compared to other surveys, a relatively high proportion of over 37 percent of those surveyed stated that they were already involved in AI-based solutions in their company. In contrast, however, there are more than half of those surveyed who find the topic of AI exciting but have not yet had the time or opportunity to deal with it.
Overall, the answers show high expectations of AI and its benefits for products and services. 60Percent of the respondents express themselves accordingly. Only 7 percent state that AI is overrated. Most of those surveyed agreed that AI should be used to support and not replace people in production.

Assessment of practical use difficult

The number of providers and solutions for the implementation of AI projects in production is growing steadily; the market is becoming increasingly confusing for users. What makes the selection even more difficult: Often several AI solutions have to be integrated into hardware components, such as machine controls. This complexity is reflected in the classification of solutions: the most common obstacle in identifying suitable AI solutions, at 65 percent, was problems in evaluating the economic benefit of an AI solution for their own application context.
At 64 percent, in second place is the question of whether the respective AI solution can be used at all in its own context. “In order to offer companies a faster overview of available AI solutions and their applications, IIP-Ecosphere is developing an AI solution catalogue, which addresses the problems of those surveyed and includes aspects of benefit and applicability,”

Legal uncertainty on the subject of data sharing

The survey results show that over 90 percent of companies already collect production data. However, almost half of those surveyed say that collecting the data needed for AI solutions poses problems for their company. There is therefore a clear need to catch up when it comes to the needs-based collection of AI-relevant data.
A mixed picture emerges when it comes to data sharing: On the one hand, 57 percent of those surveyed think that the Other companies’ data could benefit, but only 16 percent would purchase non-company data. 59 percent still see a need for clarification on the legal issues. “The responses of those surveyed also make it clear how necessary it is to understand the legal regulations in order to be able to clarify the framework conditions for the use of external data,” says Hans-Dieter Schmees, project manager for technology and standardization at the VDW.
Almost half of those surveyed stated that they use cloud solutions to handle internal company data. Interestingly, however, around two thirds of the companies also agree that production data must not leave the company. Just under 10 percent of those surveyed who commented on cloud solutions primarily rely on an on-site solution and would not use a cloud solution.

IIoT platform mainly used by larger companies

Industrial production is increasingly using IIoT platforms that support the coordinated control of machines and the centralized collection of data. Almost a third of the companies surveyed are already using such a software solution, and almost 45 percent have not planned to use it. Almost 7 percent of the companies that use a platform use their own development. According to survey results, however, it is primarily larger companies that are already actively using IIoT platforms. dr However, Holger Eichelberger from the Software Systems Engineering working group at the University of Hildesheim estimates a trend towards the increased use of IIoT platforms, which is also confirmed by the survey: “In the medium term it can be expected that around half of those surveyed Companies from large corporations to SMEs will soon be using such software.”
An obstacle to the use of AI, which was mentioned in particular in the free comments, is the age of the company’s own production machines, with which the necessary data cannot be recorded or can only be recorded with great effort. Concerns are therefore arising, especially among SMEs, that they could miss out on Industry 4.0 and AI. In addition, some companies are afraid that research and business development will lose sight of them or that they have no opportunities to (help) shape relevant developments. “Cooperation in a spirit of partnership across disciplines is therefore essential for future clout and for securing a competitive advantage, at least in the classic, economically strong disciplines,” says Dr. Alexander Broos, head of research and technology at the VDW.
The open ecosystem approach of IIP-Ecosphere with many participation opportunities and offers offers companies the chances and opportunities they need to tackle challenges together and to develop technological alternatives that make AI accessible to as many companies as possible.