During the Cold War, the U.S. government attempted to replace its staff translators with computers. The goal was to speed up the translation of important documents from Russian to English. These simple, rule-based expert systems were able to translate sentences word for word, but usually failed to provide comprehensible context and meaning for those reading them. After years of trying to improve the technology, they eventually gave up and started using human translators again.

With improved hardware and the introduction of machine learning, AI technology in the 90s could not only sort emails, but also defeat chess masters. During the 2000s, this development was also important for facial recognition, voice recognition, and, not least, better translation programmes. The picture below illustrates the relationship between artificial intelligence, machine learning, deep learning, and the development over time - from simple chess programs to deep learning in neural networks.

Illustrasjon: Relasjonen mellom kunstig intelligens, maskinlæring og dyplæring og utviklingen
over tid - fra enkle sjakkprogram til dyplæring i nevrale nettverk.
Fra Kunstig intelligens og personvernLinks to an external site.. (s.5), av Datatilsynet, 2018.

Illustration: Datatilsynet. (2018). Kunstig intelligens og personvern. p. 5. datatilsynet.no/

Over the past decade, AI, including language technology, has come much further thanks to so-called deep learning. It involves a type of machine learning that apparently makes AI less rule-bound and, at the same time, more creative. We now have advanced computer programs, so-called language models, that use deep learning to generate new texts and provide perfectly acceptable translations. Open AIs ChatGPT, Google Gemini, and Microsoft Copilot are examples of chatbots with advanced language models.

Clip 2 from AI Art: How to see like a machine

NB: the video loops after one minute

Video clip 2: Museum of Modern Art (MoMA). (2023). AI Art: How artists are using and confronting machine learning | how to see like a machine. youtube.com

Social media News media Academic literature Artificial intelligence