Real Versus Robot: How to Make Sure You’ve Hired a Human

My colleagues recently wrote about the problems AI deepfakes pose in the workplace and some current legislation addressing the risks that AI poses through exploitative images on social media and in the workplace. It is getting harder to distinguish between real and AI and this spells trouble for employers and employees. Trust is important for healthy employment relationships and AI is making trust-building in our digital world increasingly difficult.

The media is rife with examples of AI run amok, from influencing politics to faux celebrity postings. But what has received less notice are actual instances of employer bot-hiring in our virtual world. The flip side of this problem is the growing incidence of scam job postings that trick candidates into sharing personal details that are harvested for malicious purposes. Clearly, bad actors have leveraged AI to their criminal benefit and, as a result, it might be time for employers to revisit their hiring process to make sure candidates are who they say they are. Here are a few simple suggestions to consider for safeguarding candidate authenticity.

A personal meeting with a candidate is, of course an ideal (though not foolproof) way to confirm someone is who they say they are. If an in-person meeting is not possible, a video meeting is a decent substitute. While presumably technology could allow for impersonation in a video meeting, it would be more difficult for a bot-impersonator to answer interview questions on the fly. And if a candidate is reluctant to appear on camera due to apparent technology glitches, consider that a potential red flag and simply reschedule the interview for a more technically convenient time.

Vetting of candidate credentials should be undertaken before setting up interviews. Vetting might include cursory activities, such as confirming prior employers from a resume (are the employers real and do they have a verifiable history?) Checking degrees from educational institutions, another area rife with fraudulent credentials, is another basic tool. Using a bot to do those spot checks? Keep in mind that bots do not insulate employers from employment claims and while some AI programs might actually be helpful here, there is a growing legislative interest in mandatory disclosures relating to AI in the workplace (like California’s proposed “No Robo Bosses Act”), the application of which should be considered before that technology is deployed.

After a conditional offer of employment is extended (and all the necessary consents are collected), then a robust background check process can take place, with full accountability by a properly credentialed and experienced agency. That process, if undertaken properly, should promptly disclose a deepfake.

Finally, thoughtful training of the individuals who run an employer’s recruiting function (or hiring an outside professional, such as a recruiter, whose job is vetting and qualifying candidates) should be the first line of defense against rooting out real people from bots. The employment application function is becoming increasingly mechanized. But that doesn’t give employers a free pass from their responsibility to make sure the candidates they hire are authentically who they say they are.