FINAL PROJECT: Abstract and Reader's Reponse > Understanding the Importance of a Friendly Artificial Superintelligence

ABSTRACT: This article focuses on bringing the reader’s attention to the imminence of superintelligent AI and the possible catastrophic consequences if it isn’t properly controlled. The piece begins by outlining the various possible existential threats looming on humanity’s horizon, and presents the claim that superintelligent AI is both one of them and a possible solution to the other threats. The writer then explains why AI is actually more imminent than we might think. Our sheer processing ability is brought up, especially in comparison to the human brain, as well as the approaches that we’re taking that could each lead to a superintelligent AI. The classes of AI are explained, as well as why we don’t actually need to get to superintelligence before an AI undergoes an event known as an ‘intelligence explosion’. The article then emphasizes just how powerful such an AI would be, and goes over why a superintelligent AI would very likely end up ending humanity, or at least humanity’s agency, and details why some commonly suggested solutions don’t work. This moves into more promising paths of inquiry, still pointing out potential pitfalls but giving a sense of how to even start thinking about possible solutions. The articles finishes with an emphasis on how important it is to focus on this problem and solve it in time, and provides some references for further reading.

WC = 225

READER’S PROFILE: A possible reader might be rather skeptical of the severity being implied here.

READER’S RESPONSE: Well, this is something. On the same level as nuclear war and global warming? Sounds to me like he might be overstating it. Really extensive lead-up, and I’m not convinced Moore’s Law is something to really bring up, but he is being backed up by Stephen Hawking and Elon Musk… probably something to look into, at any rate. What am I supposed to do about it, anyway? Sounds like it’s out of my hands. I should probably vote against referendums on militarizing AI, I guess. I can't afford to donate to every cause that grabs my attention.
May 8, 2016 | Unregistered CommenterDD
D -- just for reference I do not think that military drone policy will be the subject of referendum votes, but I could be wrong.

Did the Rand information help?

What headings are you using to make ploughing through this document easier for the reader?
May 8, 2016 | Registered CommenterMarybeth Shea