What is artificial intelligence (AI)?
Definition of AI
Broadly speaking, artificial intelligence (AI) is the ability of a computer or program to match human behavior or thinking and perform human tasks. Instead of being programmed for every task, AI can find answers and solve problems on its own. AI employs methods and technologies such as machine learning, cognitive modeling, pattern recognition, and image and natural language processing. AI systems can be used to solve tasks that a human cannot oversee and solve due to their complexity. AI systems are versatile and accomplish different tasks. However, from a scientific point of view, the term AI is as fuzzy as the definition of human intelligence.
Weak vs. strong AI
In principle, we distinguish first between weak artificial intelligence and strong artificial intelligence.
- Weak AI targets a specific task where analyzing large amounts of data assists in problem solving. All AI systems that we know today are based on weak AI. They support us in specific use cases.
- In contrast, strong AI is defined by self-learning technologies that are used to solve arbitrary tasks, at the level of human intelligence. Strong AI in a futuristic scenario would actually develop its own consciousness and be able to creatively solve tasks of all kinds – but we are still a long way off from that today ... which is a good thing!
- 1950: Alan Turing develops the Turing test to see if a machine is perceived as intelligent.
- 1956: Scientific conference – the first time that simulated machines are referred to as "artificial intelligence".
- 1966: First Chatbot "ELIZA" is developed.
- 1972:MYCIN – Artificial intelligence is applied to mainstream medicine.
- 1997:Deep Blue – AI-based chess machine beats the world chess champion.
- 2011: Artificial intelligence is omnipresent – voice assistants are integrated into smartphones.
- 2023:ChatGPT revolutionizes the application field of chatbots.
- 20xx: Today, we can't quite imagine what AI will be able to do in the future.
The story of artificial intelligence (AI) is far from finished.
Where does artificial intelligence (AI) begin?
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.
How can you recognize artificial intelligence (AI)?
Unlike conventionally programmed applications, a self-learning AI system improves with each new task. While the programmed task will always output the same result, AI's result may vary from time to time – ideally, the result will continuously get better. This is due to the fact that AI does not follow rigid rules, but creates its own rules based on historical data and constantly optimizes itself.
The following explanation, including infographics, will provide you with detailed information about the differences between the use of artificial intelligence and the conventional approach or classic programming.
Classic programming
So far, all rules and influencing factors had to be known in order to implement them specifically for each application. Hard-coded programs calculate a result from input data. Use cases are programmed in applications, which is rather inflexible, since all interrelations and special cases must be known at the time the program is created. Therefore, classical programming is mainly suitable for clearly defined use cases.
Artificial intelligence / machine learning
Artificial intelligence is completely different. Recorded input data and the results of previous calculations are supplied. AI detects correlations and creates models that emulate system behavior. AI learns from historical data (machine learning) and can use this knowledge later on. This makes AI much more flexible, because not every special case needs to be known at the time of creation and AI figures out the correlations by itself.
Classic programming vs. artificial intelligence (AI) / machine learning Where is artificial intelligence (AI) used?
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.
How does artificial intelligence (AI) fit into everyday work?
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.
How is artificial intelligence (AI) shaping the world of production?
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.
What AI methods are commonly used in manufacturing?
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.
Will artificial intelligence (AI) replace humans in the workplace?
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.