A.I. Guardrails

Congress is beginning to consider legislation to put some guardrails around the burgeoning Artificial Intelligence technology now overtaking the U.S. and the world. Yet they seem at a loss of exactly what to do.

Pilgrim, I have a few suggestions for these contraints … and they involve indelible watermarks and a new type of copyright.

First the watermarks:

All artificially generated text, still pictures, videos and voice prints need to be watermarked in such a way that they cannot be easily altered by technology overrides. And any originating product that violates these terms will suffer severe legal consequences … and ditto for any product that overrides or reverses these watermarks,

Here are some possible watermarks:

Text — shown in reverse italics or in an obvious and unique font

Pictures and videos — a visible red or red-and-white checkerboard picture frame … or some visible watermark imbedded in the image(s)

Voice — a soft beep-beep in the background say every 5 seconds or so

And for the new special form of copyright:

All artificially generated AI text, picture or video outputs are restricted from being copied in any manner whatsoever …without identifying and specifying this special copyright and its source protection … again with severe legal penalties for violations. (Maybe calling it a “copywrong?”)

STAND UP FOR GUARD RAILS