Woven and ChatGPT resistance
Is Woven ChatGPT resistant?
In case you haven't seen it yet, soon after ChatGPT's release we did a series of tests on our own scenarios to test their ChatGPT resistance and published the results here.
The long and short of it is, some of our scenarios are more resistant to ChatGPT than others. Here are a few things we're doing about that:
- With new scenarios, we're starting to add guidance in the prep materials as to when candidates should avoid using ChatGPT. We're currently building new versions of our Architecture Debugging scenarios, and we'll include that guidance there. It certainly won't stop everyone from doing it, but we anticipate that it'll stop some folks.
- Our scorers are already catching candidates who have obviously used ChatGPT based on just recognizing response similarity and verbiage. We'll continue to flag those responses.
- We're leaning into form factors like our code review scenario to give Woven customers more ChatGPT-resistant options.
We always recommend asking a couple of follow-up questions at the beginning of your next interview with the candidate, just to confirm that the work they did on Woven was their own.
Do you have any concerns about specific scenarios that you're currently using? We are always happy to explore our library of options together.
Reach out to your CSM in your Woven Slack channel or email us at: help@woventeams.com
