The Stanford One Hundred Years Study on Artificial Intelligence provides a broad insight on how Artificial Intelligence (AI) has changed eight domains of the life of habitants of big cities in the past 15 years and how it is likely to change it in the next years, with special focus on the next 15. However, some aspects discussed in the study are worth to challenge. Here I challenge two points: first, and more broadly, that the enhancements in many domains may not be affordable to all city habitants and how it would be possible to lower such costs and second, and more specifically that the advances in education at the high school level may neither be cost effective nor necessarily positive.
Regarding the affordability of AI developed products, in the past 15 years there have been two big tendencies observed: one is that most personal technologies are highly expensive (and probably not affordable to the lower classes of society) and another is that funders have underinvested in AI research lacking some commercial application. In face of the involved costs, the development of certain technologies main not benefit everyone as the Stanford Study claims, they in fact may serve to widen existing inequalities in many aspects of a society. For example, many of the technologies mentioned by the study will be available through the use of Apps in smartphones. The price for such devices runs around 700.00 USD while minimum wage in the state of Wisconsin is 7.25 USD/hour. When performing simple math, we conclude that a worker in the state of Wisconsin would have to commit virtually 100 labor hours, (other than all other living expenses he/she has) to afford buying such a product. This is just an example in terms of personal devices but many other AI developments proposed by Stanford will involve even higher costs. The high cost involved with AI developments also extend to other domains. For example, the changes proposed in transportation will be cost dependent and probably not implemented until cost is low enough. In the domain of robotics, the slow growth mentioned by the study could also be due to sales price. Even the implementations proposed for low resource communities will depend on limited government resources and would benefit from lower cost. The main question only superficially addressed by the review study is how to lower those costs making AI implementations feasible? In my point of view, currently most developments are proposed by a few big companies and government stimulus and regulations will become essential to support smaller startups and incentivize free market competition.
Regarding the advances of AI in Education, especially at the high school level, there is no scientific evidence that robots will improve learning in face of the high costs they would impose to the system.  At the same time that individualized learning can bring benefits, it also limits social interactions, impairing the development of social skills. If such limitation is imposed or facilitated, especially at a young age, it could have catastrophic results in the personal life and career of kids, particularly the ones that already present introverted behaviors. Recent studies also claim that high exposition to technology at young ages may cause neurological side effects. Since those side effects are yet not well known, I believe that the implementation of some technologies should be avoided. 
Finally, AI developments have a huge potential to improve human life in many aspects, but implementations will have to be affordable and done with caution, sustained on supporting research to avoid unwanted consequences.
