    I believe Section II: AI by Domain does a great job of explaining what has already been accomplished with AI and just as good of a job in predicating where AI will be in the coming years. While the description of the technology is great, the ideas of how to change public policy and what needs to be addressed when considering changes to AI predicting Section III is not equally as progressive as the changes that Section II predicts. Section II predicts a cultural shift in how AI is adopted, while Section III frames the needs public policy within the current public policy climate. For the changes in culture caused by AI to be as significant as they have the potential to be, the public policy and regulations that govern them need to change just as progressively.

    AI policy, Now and in the Future, Policy and Legal Considerations contains a laundry list of legal or policy items that are based on the fears of the present day.  AI policy cannot be restrictive, a cultural shift is necessary. To see how mild the considerations are for policy, look first at privacy. The section on privacy suggests that society will still have the same attitudes towards machines and technology storing personal information. Out of both necessity and changes in generational attitudes, the requirement and need for privacy will dissipate. There is an inherent need for AI of the future to store personal data. For example, if one was taking a trip in a self-driving car, the destination and time and other metadata of the ride must be stored for both billing, safety, and liability reasons. Even though there will be a significant rise in the sensitive and metadata stored on people, there is and will be a rise in the acceptance in sharing personal data. This can already be seen today. Older generations such as baby boomers and Gen Xers are typically more concerned about what they share with websites and other technology, while millennials broadly are willing to give out sensitive information to any source and often post invasively personal life details on social media for the world to see. Whether the influx in data is a good thing for society and its people is not the point of this paper, but rather that it is happening. This change is happening and public policy will be dictated by AI, rather than the other way around.

    Another possibly problematic section of AI policy, Now and in the Future, Policy and Legal Considerations is Liability, both civil and criminal. The study makes the assumption that there will even be liability considerations in place eventually. While there will be liability laws during the early days of AI, as time progresses they will become irrelevant and unnecessary. As AI nears perfection, the incidence of liability will nearly vanish and when unfortunate or damaging events occur they will be treated as a coincidence rather than an event where fault must be found. A machine was at fault and society will someday adapt to accept that the event was a freak accident, and maybe take the event into account with reinforcement learning.

Above all else, public policy makers want the economic benefits from AI and more broadly technology, so they will accept what consequences that AI brings. AI is the future of everything, and policy will be written and regulations cut to speed its development (see self-driving cars in NV) rather than public policy written to hinder its development. 

