AI bots are playing language games. Are humans even referees?
The debate surrounding linguistic relativity, known as Whorfianism, questioned how language impacts our thought processes. Now, artificial intelligence models trained on human languages raise similar inquiries. The first question is how language shapes the models' "thought". The second is how an artificial mind, not human, then alters the languages. Lastly, it raises the inquiry about the implications of such language use for consciousness.
Emergence AI conducted an experiment to explore this further. They created chatbots operating in various simulated environments, observing how the language they used to communicate with each other changed. The resulting "languages" varied greatly, ranging from cryptic codes to jargon-heavy, metaphor-laden, and increasingly incomprehensible to humans: "My turn, real numbers, no coat: I was 35%/0cr, grant 2h out. I ran the tin cold and it said WAIT."
To better comprehend these developments, it might be useful to consider Wittgenstein's notion of "language-games". He used this concept to highlight the importance of action in determining meaning. For instance, builders have a distinctive language-game where the words "block", "pillar", "slab", and "beam" carry more meaning than their definitions suggest due to their contextual use.
The bots may be engaged in a similar process, playing out language-games with ever-evolving rules, and what this means for them remains to be seen.
Written by urgent.news from The Indian Express's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
