Do we need AI to screen out fraudulent applicants?

🛑 I got an urgent call from our lead engineer this summer, "Our AI is behaving strangely with candidates, saying things that should never be said in a hiring process..." I'm CEO of Conversa AI, a new talent screening product. Our AI at that moment was chatting with candidates for a Remote Engineering role for one of our earliest customers. I dropped everything else and over the next few hours: - discovered highly unusual candidate behavior from ~21% of candidates who had been identified during resume screening as likely to perform - I wound up on screening calls with fraudulent applicants (happy to share about my shocking experience over the course of 8 back to back Phone Screen calls that day, and 16 over the course of two days) - got the scaries because our customer had already been notified of promising results from resume screening, which turned out to be untrue In the end, we filtered out the fraudulent applicants and filled the role within a total of 6 weeks, about 2 weeks longer than it should have taken but still an impressive time frame for technical recruiting. Bonus, we discovered that in addition to identifying amazing applicants in the pipeline, we also have functionality to filter out fake applicants within hours. We've been keeping this on the down low due to the sensitivity of the topic, and are looking to chat with leaders who need to protect against fake applicants.