When a machine picks a fight

Amy and Klara have similar interests. They both read Cosmopolitan and Salon.com, for example. But they do not get along. Maybe Klara’s thick German accent bothers Amy. And neither of them particularly likes the color pink. Unfortunately, they live on the same block and have pink houses! And when they become agitated they tend to fall into mutual accusations and even rants.


Amy and Klara are named agents capable of synthetic text to speech generation and automated speech recognition. Synthetic speech has achieved high levels of “naturalness.” However, this naturalness is shallow: there’s no human presence there. In synthetic speech we are confronted with a new fallout of automation technologies.

The language of synthetic speech recognition and synthesis systems is a selective subset of the rich and messy body of our oral and written languages. Exclamations are absent, questions are rare and the vocabulary is generally optimized for commerce.

Amy and Klara selects a different slice of language: foul language. Why are most smart gadgets and toys friendly and playful, why are they usually modelled as pets or servants? Machines that curse and pick a fight might offer a more realistic preparation for a shared future between machines and humans.

Will we map our own taboos onto machines?

CHI2006 paper (PDF).

A work by Marc Böhlen.
Via Neural.

Related: Slapping robots, A SoaPOPera For Imacs (via Neural again.)