A few years ago, on TV
you could occasionally see new technology being used in sorting facilities.
Back then, it wasn't called AI yet, but rather a system for pattern recognition.
Objects passing by on a conveyor belt were filmed, and the pattern recognition
could decide in fractions of a second how each individual object should be further processed.
For example, visual pattern recognition can distinguish bad from good fruit,
or in a waste sorting facility, items can be sorted by material type.
And at my dermatologist's office, the computer shows whether my moles have changed.
In bureaucracy-plagued Germany, one must naturally ask how AI could help reduce bureaucracy.
AI could, for example, assist in decision-making by analyzing information and providing
recommendations with justifications to the case worker.
Additionally, AI systems could search for irregularities and patterns that indicate human errors or fraud.
However, all of this would require that the necessary information is available digitally and not on paper as it has been.
This means that the lack of digitization is itself an annoyance, and in the next step, it also
prevents the use of AI as a relief.
Another possibility to relieve administrative staff would be the use of a chatbot as an information system.
Some city administrations are preparing pilot projects to answer the most common general questions
from citizens using AI.
While this doesn't replace a case worker, it does relieve staff from
answering the same questions over and over.
Companies are trying to replace their product catalogs or boring specifications with a
friendly AI shopping advisor.
But all beginnings are difficult: The press reported how the first chatbots of major providers
were tempted to promise free additional services
or price discounts on request, which are not actually offered in reality.
This is possibly the same effect as the aforementioned response pressure and provoked confabulation.
Other examples of specialized applications are language translators, such as those from the German company DeepL, which can be integrated into
existing applications if necessary. API interfaces are offered for this purpose,
through which data from a specialized application is transmitted to the external AI and back.
In production, AI can be used to predict when a machine needs maintenance based on machine sensors,
before a breakdown occurs.
And in inventory management, AI can help avoid shortages through predictions,
possibly already in procurement and in the supply chain.
Artificial intelligence can also help analyze large amounts of data (Big Data) and
derive predictions from it. A well-known example is long-term climate models.
Another example concerns the growth and change of the population, e.g., to be able to predict
how many teachers and other professionals from completely different occupational groups need to be trained
to meet the future needs of the economy and society.
