At Quester, we’re built on stories – and we protect them
We’ve been talking about AI for years – the tools we leverage for interviewing and analytics are built on AI aimed at understanding, and they’re incredibly compelling and useful. However, as AI has advanced to the point where people are asking ChatGPT to do their jobs and write poems about their dogs, AI is more visibly present everywhere.
And now we have to embrace it, while we also combat it. A little.
Because the AI that’s top of mind now seems wild and futuristic, and like a game. And while it’s objectively amazing, now we’re not just getting questions about our tools, we’re getting questions about our data – and whether our stories have been compromised by people using AI to game our systems (and everyone else’s – and don’t let anyone tell you differently).
So the first thing for us to talk about is what we actually do. Our tools aren’t built to respond, they’re built to listen. Not to give answers, to be clever, or to generate information. We’re built on psychology, not just algorithms, with the very simple goal of helping people become better storytellers. This is also why it’s critically important to us to ensure the integrity of our moderated stories in the world of AI.
Our approach actually helps us make sure that our stories are coming from people. We’re designed to capture people expressing experiences and emotions, so it’s easier for us to flag what’s genuine. For example, people respond with messy thoughts and feelings – assisted responses tend toward over-elaboration and lists.
At the beginning, our AI moderator includes quality standards to catch and flag any responses that don’t meet our linguistic requirements. From there, we employ an intensive, two-part data-cleaning process. The first, an automated data cleaner, flags word count and potential fraud violations. The second is a human linguistic review, both within and across responses, that evaluates answers for quality, but also for known AI-assisted flags like answers that are too long, colons and lists, overly formal language, weird pronoun usage, or repeated (but minimally revised) responses.
We are AI believers and story advocates, so we will continue to grow and evolve all of our tools to ensure that we can always capture and leverage authentic stories, and the voices of people in our work.
