When it comes to keeping up with the latest breakthroughs in AI and how it is being used to create new types of fraud, it can be hard to keep up. Every day is a school day and if you are not learning, then you are falling behind, quickly!
We are fast headed for the unknown unknowns, so much so that knowing what to watch out for is becoming more of a complex task.
During a hiring process for a senior development engineer, we witnessed firsthand AI deception to a professional standard we’ve never seen before. Thankfully, using strategic questioning we were able to ascertain that something just wasn’t right and as we dug further, we uncovered more and more deception.
We found a network of LinkedIn profiles for people that were all fake, the further we dug, the more we saw. Candidates using fake LinkedIn pages had applied for the role and used advanced AI to answer questions in the interviews in real-time. These candidates were very good at deceiving and very quick witted. They were able to provide fake passport images and false references.
Just how many people had been duped by this masterful scam and how far-reaching is this deception? If skyrocketing global fraud is anything to go by, then this may be another type of fraud on a meteoric rise.
As with all new things encountered, there is always something new to learn.
As we dug through the LinkedIn profiles, some common characteristics became clear. Some profiles had lots of followers but no posts, some with very few connections, no profile image, some had 500+ connections but no posts; surely someone who has that many connections would be posting often. And so, the list went on and so I learned what to look out for.
It’s no longer enough to try and observe if a candidate is reading from a screen, when the video is manipulated using AI and the candidate is not the person answering the questions. An out of sync conversation can easily be passed off as a time lag on the video, this is where a strategic line of questioning is key to uncovering deceptive behaviour.
AI is creating new opportunities for fraud on a daily basis, and we are rising to these challenges and keeping our customers safe in an ever-changing world. Last year alone APP fraud in Ireland alone topped 18 million euros.
At Sedicii, we are fighting AI fraud with our own advanced technologies in the form of Zero-Knowledge Proofs and other Cryptographic protocols. Our newest offering – Confirmation of Payee gives our users secure, cross-border Confirmation of UK payee accounts for Irish businesses.
How to Spot AI-Like Answers in Job Applicant Interviews
- Overly polished or generic responses: AI tends to produce responses that sound too perfect, rehearsed, or formulaic. Look for answers that are overly general or lack personal flair.
- Lack of emotion or detail: AI struggles to convey genuine emotion or intricate details. Real candidates will add small, personal touches that reflect their unique experiences.
- Unnatural or overly structured answers: AI often delivers answers in a structured, almost robotic manner without natural pauses, hesitations, or the storytelling quality that real people offer when speaking from experience.
- Vague or abstract answers: When asked for specific examples, AI might give broad, theoretical answers instead of concrete situations involving real challenges, actions, and outcomes.
Tips to spot AI-generated fake LinkedIn profiles among job applicants
- 500+ connections or low connections
- Followers but no posts
- No recommendations
- No skill endorsements
- No verification on account
- V similar banner under name in in many profile pages
- Posting educational milestones long after the fact