Marián Šimko for DenníkN: Why we shouldn’t ask AI sensitive questions

Artificial intelligence is making significant strides in understanding natural language, but there are still many challenges to be overcome. One such challenge is teaching AI to understand the nuances and different interpretations of language, especially in the case of sensitive topics such as politics and religion.

Marián Šimko is a KInIT researcher, who is teaching AI to understand Slovak language and context. Read the interview for DenníkN (in Slovak language), where he talked about: 

  • why it is easier to teach artificial intelligence English language than Slovak;
  • how to eliminate the artificial intelligence biases;
  • how to ensure that artificial intelligence is not trained on our sensitive data;
  • whether social networking algorithms should be regulated by law;
  • how the training of AI models impacts  our environment.