Impromptu: Amplifying Our Humanity Through AI the same words should be reverberating off the walls of each Bridge nursery. It gets worse: the tablet tracks teachers’ finger strokes to see whether and how fast they scroll to the end of the lesson. And yet, this particular form of standardization seems to have helped students. Nobel laureate Michael Kremer conducted randomized, controlled trials comparing the model to local public schools, and found that, on average, Bridge students picked up more than three years of learning in only two years’ time. “The test score effects in this study are among the largest observed in the international education literature,” Kremer and his colleagues commented. Even with this seemingly disempowering technology, the key to success turns out to have been enthusiastic engagement from teachers. By one critical measure, teacher motivation soared under this system, now renamed NewGlobe: absenteeism plummeted to less than 1 percent, compared with (as above) about 45 percent in nearby Kenyan public schools—which paid significantly higher salaries. Zeambo Davis, a Bridge teacher, told a reporter from Quartz that the precise script “lets you talk less and engage students more.” Despite these successes, Bridge’s financial and business model failed. Rather than growing from 100,000 students to ten million, per the plan, the company has exited business lines and shifted its model to licensing its ideas to governments. Yet as we think of tools-plus-teachers helping as many as 600 million children worldwide who can’t currently expect to get any real education, the possibilities of LLMs on top of the New- Globe tablet model become exciting. 42