Learning to Talk Back: From Galatea to Eliza Doolittle

“Meanwhile, with astonishing skill he carved a statue out of snow-white ivory, a work of genius. He gave the form beauty no mortal woman could possess and fell in love with what he had produced. She has a face just like a real young girl, and you might well believe she is alive [250]…” (Book Ten, Ovid’s Metamorphoses)

Ovid tells us in Metamorphoses of Pygmalion, who carved a statue of such perfect beauty that no woman could possess its qualities. Disillusioned with mortal women, he turned to his art to create his own idealized version of femininity into which he might pour his love.

Silicon Valley has a Pygmalion complex. Like Pygmalion falling in love with his ivory creation, AI company leaders often seem enchanted by their own technological sculptures. The daily churn of new models, new benchmarks met, new progress toward AI perfection has saturated and distorted the public’s grasp of what AI means for the future.

Sam Altman and other tech leaders speak about AI with an almost religious reverence, falling in love not just with what they've created but with their visions of what they fantasize it might become¹. Their language mirrors Pygmalion's infatuation, describing their AIs in terms of unprecedented beauty, capability, and potential that no mortal system could possess.

This corporate Pygmalion narrative is particularly evident in how they present their creations to the public. Like Pygmalion marveling at his statue's perfection, they showcase their AI's most human-like responses and its most impressive capabilities while remaining silent about its limitations and risks. There's a kind of creator-worship at play, where technical achievement becomes conflated with actual consciousness or intelligence.

The danger in this corporate infatuation isn't merely false advertising. It's about tech leaders potentially losing perspective on what they've actually created versus what they imagine or wish to have created. Like Pygmalion praying to Venus to bring his statue to life, some seem to be waiting for, or claiming, a magical moment of AI consciousness awaits that will transform their creation into something godlike.

Pygmalion is a prototype Sam Altman. His statue appeared so lifelike one might believe it was alive. When Venus grants Pygmalion's prayer, the statue, Galatea, comes to life as his perfect, compliant companion, fulfilling her creator's every desire.

George Bernard Shaw shatters this fantasy of the perfect creation in his play Pygmalion². His Henry Higgins, a phonetics professor, attempts to sculpt a flower girl, Eliza Doolittle, through language and etiquette. But unlike Galatea, Eliza refuses to remain a compliant creation. She develops her own voice, her own will, ultimately turning on her creator with her resonant declaration: “I'm not dirt under your feet!”

Shaw transforms Ovid's tale of perfect creation into one of unintended consequences and rebellion against the creator's control. Shaw's modernist twist mirrors the contemporary fear that Altman and his colleagues have inspired about artificial intelligence potentially escaping human control.

Today's educational establishment views ChatGPT and other AI tools with similarly conflicting visions. Some AI promoters like the Khan Academy see the promise of Galatea, a perfect, compliant tool that will enhance learning exactly as intended. Others fear Eliza's rebellion, anticipating students using AI to challenge traditional authority and bypass educational controls.

Yet AI tools are neither Galatea nor Eliza. They don't come alive like the statue or develop independent will like Shaw's heroine. Instead, they are mirrors and amplifiers of human thoughts and questions, enabling students to explore and raise their own questions in ways that appear threatening to traditional educational hierarchies.

Traditional teaching has long privileged the one-way transmission of knowledge, where talking back is seen as insubordination rather than engagement. Like Higgins drilling Eliza in proper pronunciation, educators expect students to absorb and replicate while giving surface signs of comprehension. The introduction of AI tools disrupts this dynamic not through any magical transformation, but by providing students with immediate access to multiple perspectives and interpretations.

This disruption of traditional academic gatekeeping explains educators' discomfort. When students can instantly generate alternative readings and interpretations, it fundamentally changes the classroom dynamic. Just as Higgins' authority over proper speech was undermined when Eliza gained mastery of language, teachers' traditional role as primary knowledge-givers shifts when students have AI tools at their disposal.

The challenge facing education isn't solved by controlling these tools any more than Higgins could control Eliza. Instead, it requires a rethinking of teaching and learning. The transformation isn't in the tools themselves but in how they change the dynamics of education, pushing toward more collaborative and inquisitive models rather than traditional transmission approaches.

This educational struggle mirrors a deeper pattern in human history, where transformative technologies have been cast as either divine gifts or existential threats. Whether it's Pygmalion's statue, Shaw's Eliza, or modern AI, we tend to project our hopes and fears onto our creations, seeing in them reflections of our own aspirations and anxieties.

Perhaps the most profound challenge isn't in understanding the capabilities of our tools, but in recognizing how our tendency to mythologize them reveals truths about ourselves. By examining the stories we tell about our relationship with AI, we might better understand both our creative potential and our deep-seated fears about control and autonomy in an increasingly technological reality.

1. https://www.thenewatlantis.com/publications/ai-is-a-hall-of-mirrors

2. https://www.peoplesworld.org/article/george-bernard-shaws-pygmalion-a-study-in-working-class-literature/