What happened to Computers will happen to Consultants!
EU AI Act Risk Master Class
Background and Context
I was speaking to middle school students last month. They contacted us because they had AI assignments, and wanted to interview someone from an AI company about them.
Halfway through the interviews, they decided that LLMs had better answers to their questions than I did, because I showed them LLM answers to their questions while I answered them myself. By the end of the interview, they were hooked. They could see the utility of LLMs for much more than the assignment and they asked me about the impact this would have on students, teachers and society in general.
Why Consultants are like Computers
I assume that my response was similar to a response from our product manager, who looks like he is old enough to have known what a real computer was, because before 1935, a computer was a person who performed arithmetic calculations. No prize for guessing what the students thought when I told them this and asked them if anyone would consider hiring a human for their next calculation instead of using a computer.
Most of you will see where this is going. LLMs are here to stay and they are getting better by the day. I don't have much to say about teachers who ban LLMs, except that they ought to consider whether teachers with LLMs can teach better than teachers without LLMs, or whether students with LLMs learn better. But I do have something to say about consultants!
As usual, let's set the context with a concrete use case from a decade ago say, a "Prepare an IT Security Management policy" assignment. Such an assignment would take a couple of months and involve human interaction, but typically most of the effort was spent preparing a Word document.
Now press fast forward and consider how you would perform a similar task today. Would you keep the human interaction and fast track the documentation scope of the work by using an LLM to prepare a high quality template first? Wouldn't you expect to pay less to the consultants? And what's this got to do with the AIA or responsible AI?
Well... if you are an organization who currently an assignment for preparing an AI strategy document or an AI policy document, consider this. One of your first steps after the documentation has been prepared, will be to create an AI registry. This is a prerequisite for another key step: risk classification. Why not fast track your process - and save even more money - by asking your consultants to help you set up a registry instead of only writing a document which prescribes and describes these two steps? This way you will also help your consultants. If all they do is write documents, they will be replaced!
The first version of this post was published on LinkedIn.