The Risk of Cultural Sterilisation by Artificial Intelligence

Posted on January 20, 2017


The current generation of AI (Artificial Intelligence), Machine Learning or Deep Learning, amongst other inappropriately overinflated terms, has drawn a focus on the simple dislocation of labour. A few have delved deeper into the murky implications of the future of this class of autonomous, semi-autonomous and or tethered digital applications (software) and started to challenge the risks they pose. I touched on the subject in my earlier blog looking at the Cyber Security implications in ‘Bots Robotic Software the next threat surface‘ and then again more specifically in ‘Digital Dementia born of Artificial Intelligence’. But even I did not dig deep enough into the implications of these digital participants in our lives.

What is becoming clear is this subject is that going to continue revealing new dimensions for us to grapple with as its subject matter enmeshes itself into our world. The legal profession certainly think so, read Time to Regulate AI in the Legal Profession’ and it was publicly debated in the UK in October 2015 under the event title ‘You. Robot! What does it mean to Regulate AI?’ , but our brief technology history has shown that regulation is not always the best first solution to a technology problem as discussed in ‘Is Regulation of Artificial Intelligence Possible?

Stepping out of the heat of any regulatory debate for now and peeling back another layer on the implications of AI’s, turns the debate to the impact this is going to have on the rich diverse make-up of our global heritage carried forward in the myriad of cultures that make up the mosaic that is humanity.  Insights into this are expressed in the principles of ‘Linguistic Relativity’ which talks about how language affects the structure of its speakers’ world view or cognition. In practical terms for example how it can impact our economic decisions . The principle is defined in its purest form that language determines thought and that linguistic categories limit and determine cognitive categories, popularly known as the ‘Sapir–Whorf’ hypothesis, or ‘Whorfianism’.

In pre-digital world terms this is vividly displayed in the rich cultural diversity of language, thought, expressions and beliefs that have driven humanities creative innovation. This is supported through academic experiments that qualify the phenomena across various languages:

  • Chinese – Is accepted as a ‘futureless language’, it uses the same phrasing to describe the events of yesterday, today and tomorrow, unlike English which distinguish between the past, present and future. (Keith Chen ‘Could your language affect your ability to save money?.)
  • Japanese – There is a lack of any linguistic cues as to the type of noun they are modifying (Hanako Yoshida and Linda Smith of Indiana University)
  • Hebrew v. Finnish – In Hebrew, gender markers are all over the place, whereas Finnish doesn’t mark gender at all, (Scientific America). A study done in the 1980s found that, thought follows suit, kids who spoke Hebrew knew their own genders a year earlier than those who grew up speaking Finnish.
  • Russian v. Zuni – The ability to distinguish between colours follows the terms in which they are described. (Keith Chen Language Working Paper)

This is just the tip of an iceberg into the very nature of our humanity and societal wealth embedded in the subtle nuanced influence of our languages and cultural uniqueness.

The language implications of our headlong rush into cognitive digital companionship starts to crystallise. The dissolution of our language diversity into mainstream adoption of English as the predominant programming, algorithm design and analytical foundation language risks rendering us creatively constrained over time. This is compounded by the predispositions and backgrounds of the architects, designers and programmers who craft the code that influences and moulds the Artificial Intelligence of tomorrow, imposing the cultural influences, choices and pre-conceptions of a few on the many.

Do we risk an end product of an AI generation that cuckolds a by-product that could lead to the generational sterilisation of our cultural individuality, creativity and innovation?

Are we thinking broadly enough and are the implications clearly understood as we release these digital adolescents into the wild?