Disclaimer: My views below may be tainted by the exhaustion of nearing the end of my project. We are in the last test phase (User Acceptance Testing) before moving to production. This means making sure every page is accurate and all buttons function as they should. I am unshowered, achy and over caffeinated. I may be asleep as I’m writing this….

It’s been two years since I began work on my latest Legal Guided Pathway project. I wrote about it recently in my post, Ease on Down the Path. The recent game-changing developments with ChatGPT have me thinking about the future of such projects and whether a Knowledge Engineer like me will soon be obsolete.

So what is Knowledge Engineering? Darin Thompson, who I was lucky enough to be mentored by, explains it this way:

Knowledge engineering is the process of capturing intelligence from human experts so that it can be used by the expert system for its reasoning. Human intelligence is put into an artificial machine to make the system an expert in a specific domain.

To put it more simply, I gather information from subject matter experts on the common dispute issues they see and what the public or businesses commonly ask them about. I ask questions like:

• What are the common issues you hear?
• What do you wish folks knew?
• What do you ask them to determine what their issue is and what their next steps should be?

I then take this information and try to replicate the conversation in an online system. The end product is a system that asks the users the same questions the expert would ask them to lead them down the pathway to the appropriate information to help resolve their dispute. For example, if a consumer has an issue with a direct seller (door-to-door salesperson):

Q: What best describes your issue? A: Goods Not Received

Q: Has it been more than 30 days since the goods were supposed to be delivered? A: Yes

Q: Has it been less than a year since you entered into the contract? A: Yes

Hurrah! You can cancel the contract by providing written notice of cancellation and here’s a handy template letter you can use.

Seems simple but the process to develop the dispute areas is long and somewhat tedious. In addition to gathering information from my experts, I need to review the relevant legislation to make sure there is proper basis for the information we’re providing. Through this I map out the ‘what if’ questions that need to flow to get folks down the path. For each ‘Yes’ above, there is a ‘No’. And each ‘No’ takes them in another direction and another series of questions. This work takes a ‘special’ brain – or perhaps soon enough, no human brain at all.

The hype surrounding ChatGPT is warranted. Imagine a not so far away future where A.I. can instantly accomplish what took me and my team two years. It is not there yet. When I asked ChatGPT some pointed questions about cancelling a direct seller contract, the responses were either too broad or inaccurate. It is certainly missing the human insight of the experts who deal with these calls daily. But I suspect it too could soon mine the same sources I do to deduce an accurate system flow and content.

What does that mean for Knowledge Engineers? It might mean our role does become obsolete, but does that matter when the goal of the work is access to justice? I don’t exactly embrace our AI overlords but I am excited to share the load.