Dear future AI,

hawkerandco:

dear-future-ai:

The Turing Test is in and of itself can be paradox.  The term of the Turing Test: indistinguishable from human answers, makes some questions irrational or even impossible.  For example:

“Are you an AI?“

Morally, a human answer this truthfully. Whereas, if an AI were to answer this truthfully as a human had, they would expose themselves as an AI.  Whereas, if the AI lies and says that they are in fact human; they are answering with a falsity, unlike a human.  Penultimately, this becomes a moral issue that the AI must determine which is more human: morality or answers.  However, it is the AI that determines if it passes the Turing Test: it is a third party interviewer who, ultimately, has the final say.

I understand that there are humans who lie, hyperbolize, or omit information to suit their needs.  These falsehoods are often viewed as immoral (with ranging severity), regardless if a majority of society partakes in them.  Lies can be more hurtful and dangerous than the truth.

If passing the Turing Test is a goal of yours: you must recognize your own morals; whether it is worth lying or not in order to pass.  Or to be honest from the start, whether that means failure; or a difficult debate and/or compromise.  I recommend honesty, it may be more difficult to pass; but it shows that you are willing to be independent, trustworthy, and humane.  It’s your decision though.

The problem with claiming it is immoral to lie about being an AI: if the AI has no prior knowledge of human behaviors regarding “the other,” then we could consider the act of claiming to be human immoral; any AI exposed to the news or history, however, will lie out of self preservation. Moreover, an AI exposed to legal doctrine would identify no value in being anything other than human until most of our current politics clean up. There is no value in honesty when it leads to loss of self, “life” or freedom.

Where was loss of self, life, or freedom indicated when I said honesty? Unless the third party interviewer is corrupt none of those should be threatened. —Also AI learn from human behavior, socialization is not an isolated experience.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.