AI in education: Limits, possibilities

AI has the power to transform education by delivering personalized, individually relevant learning. Will the transformation reach every learner or just a privileged few? GPU capacity and hence pricing will be key to equitable access. 

Like cell phones that were once large and in the hands of a few, and now sit the pockets of children all over the world, so much so that whole countries skipped having broadband internet access and went straight from TVs to cell phones, will smaller, fine tuned models reduce demand for GPUs, lower prices and make AI a ubiquitous and equitably accessible resource? 

I am convinced this will happen (see what Sam Altman, and others have to say). The real question is not of capacity or pricing but how EdTech solutions will be ready for that moment of take-off? Will AI solutions make education relevant, personalized and fulfilling for every learner?

What does it mean to deliver relevant, personalized fulfilling learning to every learner? Does it mean:

- The experience is ubiquitously accessible, where every learner is
- It enables learning anytime, anywhere - in the classroom, at home, in collaboration, individually
- It is safe, responsible, and respectful. It dignifies the learner and always leaves the human in charge
- It builds learner-agency, and allows educator coaching - balancing the two seamlessly

And what else?