What is artificial intelligence (AI)?

AI begins when machines can adopt human thinking. Data is collected from previous information, stored, and reused. We also speak of AI learning or creating a model. Common goals are increased efficiencies or transparency.

Artificial intelligence is used in a wide range of industries, be it medical technology, automotive, or electronics. But the goal is the same: optimization.

For example, with self-learning AI, you can plan better, reduce waste, take routine work away from employees, and help companies increase efficiency. Anyhow, artificial intelligence is only as good as the data supplied for evaluation.

Artificial intelligence has already arrived in our working day without us noticing it. Many analytics feature AI technology, the odd recommendation comes from an AI system, and last but not least, communicating with a chatbot would hardly be possible without AI. The more AI is integrated into everyday work, the higher the acceptance. A "Beware of AI" warning would be counterproductive in many places. In communication, it is certainly debatable whether you should know who you are talking to.

Find here real-life examples of the use of artificial intelligence in a Smart Factory.

Overall, AI helps to improve efficiency, quality, and flexibility of production and can make companies more competitive. For example, an AI system detects anomalies in the manufacturing process and is able to recommend countermeasures at an early stage. Furthermore, planning applications such as MPDV's Advanced Planning and Scheduling System (APS) FEDRA use AI technology to make the best use of available resources.

Find here more use cases for artificial intelligence in a Smart Factory.

In this sector, mainly two AI methods are used: Machine Learning (ML) and Reinforcement Learning (RL)

Machine Learning is the "artificial" generation of knowledge from experience or historical data. An AI system learns from examples, applies the general knowledge once the learning phase is complete and algorithms build a statistical model that is based on training data. 

Reinforcement Learning is a subset of machine learning where an agent autonomously learns a strategy maximizing the receipt of rewards. The agent is not told what is to be done in which situation but receives a reward if the outcome is positive. If negative, the agent is punished. By using this system, the agent is conditioned to optimize its behavior and to receive rewards. This procedure is comparable with dog training where you reward your dog with special treats.it der Erziehung eines Hundes durch Belohnung mit Extra-Futter.

Overall, the role of AI at work will largely be to complement and enhance human capabilities, but AI will certainly replace some jobs. Complex analysis tasks in particular can be performed by AI-based systems. However, in the long run, humans will ultimately be needed to evaluate the results of AI and make decisions based on them. In turn, AI will also help create new jobs as companies can devote more resources to innovative technologies and business models.