> So by lying to the machine are you not lying to the person who programmed it?
Sure. And this is a problem that has many layers of depth to it, each with many interesting ethical questions waiting to be explored. It would probably make a really good book for someone who wants to make a name for themselves in philosophy.
At the most basic (simplistic?) level though, the reason software exists is to create a mapping such that every valid set of unique inputs always produces the same output. The question is, does the programmer sacrifice their humanity in doing so?
The way a human makes decisions is by looking at all the relevant facts, and then making the best decision in each case based on the information available. Software differs from this in that:
- You make a single decision, in advance, rather than a new decision for each case.
- This decision is only based on what you thought would be important at the time the decision was made, without taking into account the unique situational factors 'at run time'.
- You know in advance that there are going to be both false positives and false negatives, and determine the acceptable rate of each based on the economics of making hundreds or millions of these decisions in aggregate. Of course you also make mistakes when making one off decisions also, so the error rate of software can either be higher or lower than with human decision making. But even if the error rate is lower, the sorts of mistakes that are made aren't necessarily the same.
- The assumptions made by any given algorithm may become increasingly outdated over time, and often stick around long after any given individual would have stopped making the same decisions.
There are more differences obviously. And of course there are different types of software; lying to someone over a phone on the basis that there is some software converting the conversation from analog to digital and back again is probably different than telling some porn site you're 18. But overall I do believe that this may be one of the most important ethical issues of our time in the coming century.
Sure. And this is a problem that has many layers of depth to it, each with many interesting ethical questions waiting to be explored. It would probably make a really good book for someone who wants to make a name for themselves in philosophy.
At the most basic (simplistic?) level though, the reason software exists is to create a mapping such that every valid set of unique inputs always produces the same output. The question is, does the programmer sacrifice their humanity in doing so?
The way a human makes decisions is by looking at all the relevant facts, and then making the best decision in each case based on the information available. Software differs from this in that:
- You make a single decision, in advance, rather than a new decision for each case.
- This decision is only based on what you thought would be important at the time the decision was made, without taking into account the unique situational factors 'at run time'.
- You know in advance that there are going to be both false positives and false negatives, and determine the acceptable rate of each based on the economics of making hundreds or millions of these decisions in aggregate. Of course you also make mistakes when making one off decisions also, so the error rate of software can either be higher or lower than with human decision making. But even if the error rate is lower, the sorts of mistakes that are made aren't necessarily the same.
- The assumptions made by any given algorithm may become increasingly outdated over time, and often stick around long after any given individual would have stopped making the same decisions.
There are more differences obviously. And of course there are different types of software; lying to someone over a phone on the basis that there is some software converting the conversation from analog to digital and back again is probably different than telling some porn site you're 18. But overall I do believe that this may be one of the most important ethical issues of our time in the coming century.