Here at Ars, we’ve done plenty of coverage of the errors and inaccuracies that LLMs often introduce into their responses. Now, the BBC is trying to quantify the scale of this confabulation problem, at ...