This is an RSS Club only post! Please do not share.
We are fast approaching the end of another year and what a year it's been. I won't dwell on a lot of what's been going on but wanted to say something about a couple of threads that I feel naturally come together.
As I wrote recently, advances in generative AI mean that it is becoming increasingly difficult to tell if any individual work has been made by a person. A long-standing blog may attest to the 'author' being a real person but there is no guarantee that any specific post has not been created by an AI.
Part of blogging, at least for me, is that it fosters genuine connection with real people. This is where I see the current trend in 'blogging pen pals' intersecting.
An AI can make a decent approximation of 'content' but a 'lived' life, and all that entails, can only be communicated by a person.
There are calls to reduce the use of AIs, especially for image creation, as they are trained on public information which may be copyrighted. Take this post from Michelle Barker for example. (Via Tracy).
The argument is that even if someone will never pay an illustrator to create images, the use of generative AI normalises this process and creates an impact that goes far beyond any individual (non-)transaction.
Barker also reiterates what I was saying:
When I look at a piece of art, I'm looking at all the ways the human experience has shaped what I'm seeing
I make no secret of that fact that I use AI – primarily for questions around coding and, recently, for images – so am I just part of the problem? Or are we asking the wrong questions or need better transparency and control?
But what of pen pals?
I think the move back towards 1-2-1 communications is an extension of the existing 'shrinking of social scope' that has been happening elsewhere. People are rejecting the global conversation in favour of smaller, more focused, groups – usually private – in order to escape the noise, spam and abuse. Fears that generative AI will only further flood the streams with homogenous, potentially harmful crap is only serving to accelerate this process. As the scope of training material for LLMs is widened it becomes just a matter of time until they are recycling the same old information, feeding off the fruits of their own labour – shrinking the informational gene pool so to speak.
Striking up a 1-2-1 conversation with the aim of really getting to know the other person is the ultimate rejection of generated content and the ultimate reduction in scope. It is the logical next step in our search for humanity across the void of the internet. As I say, a lived life, full of experiences, emotion and nuance, can only be communicated by a person. 1
Housekeeping
I've fixed an issue that meant it was almost impossible to comment on an RSS Club post due to the URL being reset without the required parameter.
-
until we have truly sentient AIs but then the point will be moot ↩