The Impact of AI on Terminology Management - Interview with Barbara Karsch

For more from GALA Experts, subscribe to our YouTube Channel

Barbara Inge Karsch has been a passionate terminologist for over 15 years, initially at J.D. Edwards and Microsoft. Now, as consultant and trainer for BIK Terminology, she is assisting clients with terminology projects and teaches in the online Master’s Program at NYU. She is a US delegate to TC 37 and project lead of the revision of ISO 12616. In this GALA interview we discussed the impact of AI on terminology Management. 

Here are a few snippets of our conversation.

03:54 Impact of AI on terminology management

In 2021 when you asked us, many of us actually said that AI was benefiting from us, right? That was generally the tendency of our answers because obviously machine translation is based on AI and it has benefited from good solid terminology in source and target texts. There were a few who ventured to guess, but that was, I think, the main idea.

Last spring I wrote an article for the ATA Chronicle and I said, "well, look at us, we couldn't really predict this". Maybe I'm not the only one in the industry that got surprised as much as I did by all of a sudden having AI at my fingertips.

10:43 Potential benefits and risks of AI in the field of terminology management

If we look at the terminology process, since we're talking about localization, one of our first steps is always term extraction, term recognition.

The next task that I thought would be interesting is definition authoring, because that often takes us a long time or is one of the more time-consuming tasks. Sometimes you get a definition that sounds really good. And that is one of the risks that we have, that we cannot really judge, is it correct or not? We always have to go back and validate what Chat GPT or some other tool is giving us.

What it did extremely poorly was the next step, the context extraction. A context is a piece of text that contains the term in question that covers the concept we're working on. It has to be a verbatim quote. It must not be altered. I have not been able to get ChatGPT not to make changes to my context. So, this is definitely a warning right there.

19:17 How AI will change teaching

Well, the immediate thought in January when we first met at NYU was that we need to provide guidance on how to use AI safely. This is really going to be my focus for the next few months. 
I don't think it's hit that student population yet, but I imagine that's going to change rapidly in the next few months. And so we as instructors have to be ready to provide some guidance for that. 

22:19 What is happening in the ISO world of standards

The standards that ISO produces are very thorough and they go through a rigorous process with a lot of people from all over the world contributing. And that is just not a process that you can do overnight. 

 

With AI, I think we're going to move a little bit faster, and I've already seen the initial communications, the initial committees that have been formed or are in the process of being formed, because most of our standards will need to be updated in one way or another.

Sign up for our newsletter on globalization and localization matters.