Computer Programming - An Overview (1)
Computer Programming - An Overview (1)
This document and its content are copyright of Josh Tenenberg - © Josh Tenenberg
2022. All rights reserved. Any redistribution or reproduction of part or all of the
contents in any form is prohibited.
the kinds of programs that we will be writing in this course this is less of an issue.
But for complex programs that operate in the real world, such as programs that
recommend to judges sentencing terms for people convicted of crimes, or
programs that run fully automated (self-driving) cars, or even ordinary Internet
search engines like Google, the effects can sometimes lead to unethical, unjust, or
harmful outcomes. Why does this happen? The short answer is that, since
computer programs slavishly follow their instructions, the computer programmer
has to specify absolutely every anticipated possible world event that should be
taken into consideration by the computer in executing its program, and how the
computer should respond to these events. Alternatively, with Machine Learning
programs, the computer will have to try to learn about the world from its inputs. In
practice, whether explicitly programmed by a person or using Machine Learning,
there will inevitably be some kind of mismatch between the intended behavior of
the computer system and its actual behavior. Which is frightening in a world in
which such high-risk things as driving, operating nuclear powerplants, running
missile detection systems, and even teaching are increasingly coming under the
control of computers. We will learn more about these mismatches, the anticipated
and actual problems that they are presenting, and how human beings are
responding to this new world of computer-controlled existence to try to mitigate
the harms.
This document and its content are copyright of Josh Tenenberg - © Josh Tenenberg
2022. All rights reserved. Any redistribution or reproduction of part or all of the
contents in any form is prohibited.