People paid to train AI are outsourcing their work to AI

Re:People paid to train AI are outsourcing their work to AI by iinnPP on Thursday June 22, 2023 @07:17PM (#36440571)

The moment a detector is taken seriously is the moment it will be trivially beaten by another AI designed to beat the detector.

[ Parent | on HN ]

Re:People paid to train AI are outsourcing their work to AI by vminvsky on Friday June 23, 2023 @02:35AM (#36443562)

i would recommend u read the paper. the contribution isnt a detector thats meant to be taken seriously; but a detector that works in a very specific task. they then use this to estimate use of LLMs on MTurk

[ Parent | on HN ]

Re:People paid to train AI are outsourcing their work to AI by lisasays on Friday June 23, 2023 @10:07AM (#36447005)

Is it now?

Adversarial training isn't infinitely scalable either, has its limitations also.

Also - the moment that companies start training models to resist detectors, they expose themselves to regulation. Won't stop dark AI models running on some website somewhere, but it can be very effectively applied to companies running at Google or OpenAI scale.

[ Parent | on HN ]