Anthropomorphism is the tendency to attribute human qualities such as feelings, intentions or understanding to machines. With chatbots this comes naturally because they write fluently and personally. It can make interaction more pleasant but carries risks: people may trust answers too much or share sensitive information. Clear notices that users are dealing with AI and honest statements about its limits help counter this.
Conversational AI covers technologies that enable computers to hold dialogues in natural language, either in writing via chat or spoken via voice assistants. Modern systems are usually based on large language models that understand requests, take the conversation history into account and formulate suitable replies. Organisations use them in customer service, for booking appointments or for internal help desks. Reliable knowledge, clear limits and a handover to people when the AI cannot help are essential.
AI ethics deals with the moral questions surrounding the development and use of AI: fairness and discrimination, transparency and explainability, data protection, responsibility for errors, safety and the effects on work and society. For organisations it becomes concrete in guidelines, risk assessments, human oversight and honest communication about AI's limitations. Example: before a CV screening tool is introduced, it is checked for whether it disadvantages particular groups.
A chatbot is a program that receives questions in text or speech and answers them in a dialogue. Simple chatbots follow fixed rules and decision trees, while modern ones use large language models and understand freely worded requests. Typical uses include customer service, internal help desks and appointment requests – for example a website chat that answers questions about delivery times around the clock.
The Turing test goes back to the mathematician Alan Turing, who described it in 1950 as the “imitation game”. A person holds written conversations with a human and a machine; if they cannot reliably tell which is the machine, the test is considered passed. Today's language models often hold very human-sounding conversations, but the test only says something about convincing language, not about understanding or reliability.
