AI could choke on its own exhaust as it fills the web

Article by Ina Fried and Scott Rosenberg: “Scott RosenbergThe internet is beginning to fill up with more and more content generated by artificial intelligence rather than human beings, posing weird new dangers both to human society and to the AI programs themselves.

What’s happening: Experts estimate that AI-generated content could account for as much as 90% of information on the internet in a few years’ time, as ChatGPT, Dall-E and similar programs spill torrents of verbiage and images into online spaces.

  • That’s happening in a world that hasn’t yet figured out how to reliably label AI-generated output and differentiate it from human-created content.

The danger to human society is the now-familiar problem of information overload and degradation.

  • AI turbocharges the ability to create mountains of new content while it undermines the ability to check that material for reliability and recycles biases and errors in the data that was used to train it.
  • There’s also widespread fear that AI could undermine the jobs of people who create content today, from artists and performers to journalists, editors and publishers. The current strike by Hollywood actors and writers underlines this risk.

The danger to AI itself is newer and stranger. A raft of recent research papers have introduced a novel lexicon of potential AI disorders that are just coming into view as the technology is more widely deployed and used.

  • Model collapse” is researchers’ name for what happens to generative AI models, like OpenAI’s GPT-3 and GPT-4, when they’re trained using data produced by other AIs rather than human beings.
  • Feed a model enough of this “synthetic” data, and the quality of the AI’s answers can rapidly deteriorate, as the systems lock in on the most probable word choices and discard the “tail” choices that keep their output interesting.
  • Model Autophagy Disorder, or MAD, is how one set of researchers at Rice and Stanford universities dubbed the result of AI consuming its own products.
  • “Habsburg AI” is what another researcher earlier this year labeled the phenomenon, likening it to inbreeding: “A system that is so heavily trained on the outputs of other generative AIs that it becomes an inbred mutant, likely with exaggerated, grotesque features.”…(More)”.