Freedom of Speech and Generative AI

Photo Credit: Mirko Tobias Schaefer, CC BY 2.0 via Wikimedia Commons "https://commons.wikimedia.org/wiki/File:Bios_robotlab_writing_robot.jpg"><img width="512" alt="Bios robotlab writing robot" 

Freedom of Speech and Generative AI

I will not tangle with court opinions on what constitutes free speech for human beings. However, I have a few opinions of what anonymity should be offered to Generative AI. Suppose the law says that generative AI systems should watermark all the output they create digitally in such a manner that everyone's browser can in principle mark such material as "Produced by Generative AI". One result would be that students cannot submit such material to their teachers and claim credit as if it was their own writing. Ideally, the watermarking should identify uniquely the system that authored the material concerned. This provision should cover all output of Generative AI systems including images.

Secondly, watermarking would inform every user of the Internet as to the origin of text they rely upon – was it created by a human or a machine. 

We have enough of fake news already, and we do not need to have it produced in industrial quantity! For instance, if this safeguard was in operation now, no confusion would have been caused by a synthetic image showing a massive smoke column near the Pentagon https://www.youtube.com/watch?v=7mWXhGfTyso

In conclusion, I am not ready to confer citizenship on Generative AI systems.

Srinivasan Ramani