The best part about this story is that although the legal citations returned by the AI were in correct format, none of them nor their text existed in reality.
It's even worse. Perhaps at that moment they could have claimed the citations were in error based on a psuedo-argument that they were making for their client, make their apologies and perhaps get away with a minimal sanction. They later then submitted
faked versions of those citations, which were themselves generated with GPT - that's turning what could potentially be played off as an error of judgement into
outright fraud. I'd be amazed if these "lawyers" ever practice again, that's career ending.