Luciano Floridi and Anna Nobre have published an interesting paper on SSRN about ‘conceptual borrowing’, the process by which a new discipline develops its technical vocabulary partly by ‘borrowing’ concepts from neighbouring disciplines, often giving the concept a new meaning.

The authors argue that through this extensive conceptual borrowing, the discipline of AI has ended up ‘anthropomorphizing’ computers, as computational brains with psychological properties. The brain and cognitive sciences themselves have developed a vocabulary that has led to descriptions of the brain and mind as if it were a computer that processes information, as if it were a ‘biological computer’.

This mixing of the jargon of these disciplines can lead to confusion and even harmful assumptions and consequences.

Artificial intelligence and human intelligence are far apart, but using the vocabulary of other, seemingly closely related disciplines can give a different idea. Clarity about concepts is really important, especially when using a concept in a new context.

The same happens with concepts from archival science being used in information and computer science with a completely different meaning. And vice versa! It leads to confusing discussions where people are talking about the same thing but completely misunderstand each other because of different meanings!

Interesting paper! Essential reading if you work with AI specialists!

L. Floridi, and A. Nobre (2024). Anthropomorphising machines and computerising minds: the crosswiring of languages between Artificial Intelligence and Brain & Cognitive Sciences (Centre for Digital Ethics (CEDE) Research Paper, No, forthcoming. Published on SSRN.

Published as a LinkedIn post here in september 2024.