either you consider that lawyers have an added value from IA and you consider they can take an advice from an IA but they are able to see through its bullshit,

either you consider that lawyers have an added value from IA and you consider they can take an advice from an IA but they are able to see through its bullshit, or you consider they have little added value when provided IA input, and then once AI is good enough/better as a whole (considering cost and availability) they are of no use.