AI contracting tools raise a problem I haven’t seen discussed. Obviously, AI content-creation demands humans in the loop, but if we humans don’t do the underlying drafting or report-writing or much of it – if the technology creates content for us – we may never understand the project well enough to assess the AI’s output.

I’m not raising the usual concern about AI replacing junior staff, leading to a future w/o experts. This relates to our understanding of a specific contract or other project, not to training in general.

I often find that deep understanding of a deal or business model only arrives after hours of work creating a contract or report, particularly for a complex project. The resulting deep understanding may be my most important contribution, in some cases helping the client understand its own business. But if the AI does most of the content-creation, I may never get that sort of understanding. I’d then fail to deliver key value – and I may lack the depth needed for expert review of the AI’s draft, redline, or other output.

Am I wrong? Do you have a different understanding? I’d love to know. I’ve found this a relatively small issue so far, since I don’t find AI outputs good enough to replace more complex content-creation. They only edit. But as the models improve, this could become more serious.