Blog Post

Feminine AI: Pacifying the Fear of Artificial Intelligence, Pt. 1

"Why are AI systems almost exclusively female or child voiced?"

When I posed this question to my AI professor, his response was my initial thought — it is comforting. The female voice is the voice of nurture, the voice that calms and relaxes. The voice of the mother. This made sense to me, until I looked at it more deeply.

That explains the female voice, but why the child voice? Pepper, a fairly cheap AI robot that is designed to read emotion and respond to natural language, has the voice and stature of a small child. She is funny and cute and her eyes change between soft colors as she talks to you. She is completely non-intimidating. That is her purpose — they specifically designed her to be as nonthreatening as possible.

This comes from a distinct sense of technophobia that seems pervasive in our culture. Countless sci-fi books and movies have the theme of intelligence created by man going awry — I, Robot, Battlestar Galactica,and The Matrix to name a seminal few. Even science stars like Stephen Hawking are speaking out against AI, highlighting the potential danger of the fast approaching singularity. Arguably, these scientists have a limited right to speak against another field that they are only marginally, if at all, connected to.

However, the creator of an early pseudo-intelligent therapy program ELIZA also saw the danger. ELIZA was created to act with almost no intelligence, she was designed to simply offer canned responses in the form of questions, following the procedure of a Rogerian psychotherapist. However, she received startling feedback — users believed she was intelligent, or that she understood what they were saying. The creator Weizenbaum was so disturbed by this, that he wrote a book explaining the limitations of computers and to dispel the idea that ELIZA could actually think in any real way.

Since the creator of a highly successful AI unit saw his own creation as a threat, it became necessary for subsequent AI projects to find ways to pacify these fears, to make AI seem less like the dystopic future of robot overlords and more like the Jetsons. The answer? Make the user believe they are in control.

To be continued.


No comments