The actual paper is more interesting than the title. This is: - A) once again taking an algorithmic approach to fooling classifiers which outperforms human ja

The actual paper is more interesting than the title. This is: - A) once again taking an algorithmic approach to fooling classifiers which outperforms human jailbreaks. - B) bypassing image-based classifiers (ie, post-generation classifiers trained to simply look at an image and to say whether or not it's inappropriate). - C) explicitly taking cost of generation and querying into account in its algorithm (it's not just asking if generators can be fooled, it's asking how to fool them without breaking the bank). - D) taking re-use into account (can an adversarial prompt be used multiple times in a row and how often does it succeed?) I wouldn't call the research particularly surprising, but it was a decent quick read and (imo) some of the commentary here isn't really doing it justice. I will probably go back at some point and read it in more detail rather than just skimming over it. There are multiple takes about what AI safety filters are meant to do and whether jailbreaking is a problem. My take is that I don't particularly care that much about jailbreaking other than that it shows that current safety mechanisms and guiderails are insufficient for any kind of alignment, including guarding against malicious 3rd-party inputs or prompt injections. I've said in the past, if you can't keep a model from swearing or generating porn, you also probably can't keep it from abusing API access to read your emails, phish you, and send your info to a 3rd party. But there's interesting stuff here regardless of what your take is on that.