	Research in Artificial Intelligence is accelerating faster than ever, opening new domains for AI to 
interact. Advances in machine learning and neural networks has expanded problem sets and created 
real world applications for previously theoretical models. Technology experts like Elon Musk and Nick 
Bostrom have taken a strong stance against AI as it approaches human intelligence, but still we see Tesla 
Motors (Musks company) using machine learning to develop self-driving cars. The largest technology 
companies in the world are delving into AI to harness the power of this growing technology into their 
own systems. Sandford University developed a report last year examining AI research and projections 
for the next 15 years. The study panel responsible for the report put forth a multitude of ideas 
addressing various problems posed by the development of Artificial Intelligence, however, these are 
optimistic at best and reflect the theoretical limits of AI.
	The report in question reads more like George Orwell than a real world evaluation. Flying cars 
are briefly mentioned on page 7, preceding a comparison between personalized robots and the 
Roomba. Over the past century, technology has developed exponentially and futuristic claims followed 
but history has shown the optimism of such claims. Undoubtedly, the experts on the panel have an 
understanding of Artificial Intelligence beyond my own, but may be blind to the impact of their claims.
	For instance, the study panel defines health care as domain for AI growth. If say, a device is 
invented to prescribe and administer medication autonomously, who profits from such a device? Some 
trained professional is no longer needed, leading to an economic shift to provide fewer doctors. The 
patient could be limited in their communication with the device, so any personal concerns about side-
effects is ignored. By the definition of machine learning, the decision method for the device is beyond 
the limits of human understanding of the problem, else hospitals would rather hire a doctor. Those who 
truly profit are not directly impacted by the theoretical device. The hospital is able to save money while 
the company responsible profits economically from the sale of their product. What incentive do either 
party have to serve the populous when considering the adoption of the device? These questions are 
severely lacking in the field thought to produce such things in as little as 15 years.
	The Study Panel does create a few guidelines to aid the development of AI, but in 
implementation are unfeasible. In one instance, they posit that government officials on all levels have an 
adequate understanding of AI. If this was even possible, the education officials receive would be 
outdated compared to the emerging technologies they review. Officials will answer questions that 
cannot be conceived until the technology exists to ask such a question. How then, can one be trained to 
answer questions in such broad domains as transportation, personal robots, healthcare, education, 
public safety before they can be conceived?
	We cannot presume to know the world in 15 years and what technology will take hold. We can, 
however, reformulate our plan to develop Artificial Intelligence. Our society must develop specific laws 
to test and gradually implement AI technologies so as to not create excessively dramatic shifts in 
infrastructure or the economy that could collapse sectors of our country and lead down some 
unforeseen path. We must react, as a society, to the impacts of AI to preserve a stable society which 
supports itself, not just those with economic profit.
