Skip to main content
    KI-Beratung Deutschland
    🇩🇪DE🇬🇧EN

    AI (Artificial Intelligence)

    Definition

    Artificial Intelligence (AI) is the umbrella term for software that takes on tasks that normally require human intelligence — such as understanding language, interpreting images, making decisions or learning from experience.

    Meaning in a Business Context

    In a corporate context, AI today mainly comprises machine learning, language models, computer vision and speech processing. Classic rule-based systems — for example hard if-then logic in accounting software — are also frequently marketed as AI, but technically they are something different. For decision-makers, it is worth looking at the underlying technology: statistical learning methods adapt to data and are only as good as their training material; rule-based systems are rigid but fully predictable. Mid-market AI projects combine both worlds — for example a rule-based workflow engine, complemented by a language model for unstructured inputs.

    What are the four types of AI?

    A common classification distinguishes four types: reactive machines without memory, limited-memory systems, theory of mind and self-aware AI. In practice, only the first two exist today. All systems used in companies, including large language models, currently belong to the second category.

    The four types at a glance

    Reactive machines

    Reactive machines respond only to the current input and do not take earlier situations into account. They do not retain memories or adapt their behaviour through experience. A classic example is a chess program that evaluates possible moves in the current board position without learning from previous games.

    Limited-memory systems

    Limited-memory systems also use information from the recent past to determine their next output or action. Many learn a model from extensive training data and process the current context at runtime, without permanently storing every interaction. This includes today's language models and many applications for image recognition, forecasting or assistance.

    Theory of mind

    AI with theory of mind could take other people's beliefs, intentions, knowledge and feelings into account as distinct mental states. It would need to do more than recognise words or gestures: it would have to infer what a person knows or expects. Such capabilities are being researched, but are not yet available as generally reliable AI systems.

    Self-aware AI

    Self-aware AI would be conscious of its own existence and internal states and could reflect on them as its own experience. This goes well beyond convincing language, descriptions of itself or recognising its own limitations. Such consciousness is hypothetical; today's AI generates outputs from learned patterns and has no demonstrated subjective experience.

    For more information, see Artificial intelligence for businesses and the glossary overview.

    Related Terms

    Further reading at KBD

    AI Consulting for Companies