Clarkesworld had to shut down its submission form due to an unmanageable deluge of bad AI-generated fiction – and no, AI-detecting AI filters don’t work (for many of the same reasons, by the way, why “crime prediction” algorithms don’t).
Economics 101: if the cost of content drops and the cost of thoughtfulness doesn’t, the volume of thoughtless content rises. Good if you are in the “content” business, bad if you are publishing sci-fi (or doing science or investigative journalism). Besides, unless the demand for bad content rises exponentially (and we’re already saturated, I think and hope), none of it will lead to higher pay: your output per work unit is higher, but so is everybody else’s, and the demand doesn’t keep pace, so prices will fall —are already falling all around us— and you have to work more to produce much more to get paid, if you are lucky, more or less the same.
This isn’t an anti-AI observation. I’m convinced, I *know*, that AI can improve productivity, wages, and quality of life. But you have to deploy it to move the frontier of the possible, not to make cheaper what was already nearly oversupplied. We already had all the “content” we needed, and more – the bottleneck is in the information, thought, and knowledge, not words and images and videos. Having more maps doesn’t make your territory larger.