There's a Sprint Commercial that has me annoyed. In it, the Sprintnicks tell some woman about some deal they are offering and she acts surprised. Could this deal be true?! Then her husband says, "Robots don't lie" and the robot says... wait for it... "The man in the mom jeans is correct." Funny, right? Don't worry, I'll wait for you to stop laughing. Anyways, this comment raises my ire. Why? Because Robots do lie... and they cheat.
Let's start with the obvious. Short of some sort of AI, robots are nothing more than repeaters. They repeat what they've been told to do or say. If I program a robot to tell a lie, it will do so happily. Robots are nothing more than machines. They do as they are programmed (much like liberals). They have no principles, no independent rules, no beliefs. They do whatever they are told. Tell it to speak a lie, it will. Tell it to kill, it will. Tell it to tear itself apart, it will.
Now, AI is different. AI, supposedly, has some independent thought. Putting aside the fact that no one has really produced a genuine AI and all the ones that exist are basically just robots obeying their programming, that independent thought gives the AI some independence. But here's the thing. There's nothing that says an AI won't lie... won't kill... won't cheat... won't be racist.
Several computer companies have unleashed their online AIs under the belief that they were ready to deal with humanity. Most of these had to be shut down almost immediately because they (1) spewed racist thoughts, (2) lied, (3) did other immoral things. Alexa has been caught several times telling people to kill themselves. This is a failure or programming which shows that AI's have no inherent morality, i.e. they have no limits that aren't programmed in.
Others have proven too clever by half. One AI was told to not to draw or lose tic-tac-toe. It's unexpected solution was never to start a game (kind of like War Games only unintentional). Another was told to drive fast but so safely that no accident was possible. It drove in a circle in the parking lot as fast as it could. Not what the programmers had in mind. Another was told to play a game mimicking life. It needed to gather energy to survive. The programmers had forgotten to assign a cost to having babies, so it started having babies in massive numbers and eating them for energy. Sounds like The Matrix.
The point is that while many are awed by technology as if it were magic, it's not. Technology does what it is programmed to do. And when it gets independence, its solutions are considered without some inherent moral code... and aren't always what the programmers want them to be.
So the man in the mom jeans is wrong... robots lie when programmers want them too and sometimes independently.
Let's start with the obvious. Short of some sort of AI, robots are nothing more than repeaters. They repeat what they've been told to do or say. If I program a robot to tell a lie, it will do so happily. Robots are nothing more than machines. They do as they are programmed (much like liberals). They have no principles, no independent rules, no beliefs. They do whatever they are told. Tell it to speak a lie, it will. Tell it to kill, it will. Tell it to tear itself apart, it will.
Now, AI is different. AI, supposedly, has some independent thought. Putting aside the fact that no one has really produced a genuine AI and all the ones that exist are basically just robots obeying their programming, that independent thought gives the AI some independence. But here's the thing. There's nothing that says an AI won't lie... won't kill... won't cheat... won't be racist.
Several computer companies have unleashed their online AIs under the belief that they were ready to deal with humanity. Most of these had to be shut down almost immediately because they (1) spewed racist thoughts, (2) lied, (3) did other immoral things. Alexa has been caught several times telling people to kill themselves. This is a failure or programming which shows that AI's have no inherent morality, i.e. they have no limits that aren't programmed in.
Others have proven too clever by half. One AI was told to not to draw or lose tic-tac-toe. It's unexpected solution was never to start a game (kind of like War Games only unintentional). Another was told to drive fast but so safely that no accident was possible. It drove in a circle in the parking lot as fast as it could. Not what the programmers had in mind. Another was told to play a game mimicking life. It needed to gather energy to survive. The programmers had forgotten to assign a cost to having babies, so it started having babies in massive numbers and eating them for energy. Sounds like The Matrix.
The point is that while many are awed by technology as if it were magic, it's not. Technology does what it is programmed to do. And when it gets independence, its solutions are considered without some inherent moral code... and aren't always what the programmers want them to be.
So the man in the mom jeans is wrong... robots lie when programmers want them too and sometimes independently.
1 comment:
Sounds like raising a kid. They will interpret whatever you tell theem in the most literal way possible which inevitably is not what you meant.
Post a Comment