AI fakers exposed in tech dev recruitment: postmortem

Also it's crazy how it's not that easy to find a good job any more and yet there are these (quite obvious fakes, maybe in hindsight) who are almost getting offers over badly deepfaked video calls. Saying you're from Poland, saying you got your degree in Poland but not being able to speak Polish ?

Well, even before AI we had qualified people being paid to go to interviews for scammers. Or qualified candidates outsourcing their jobs after being hired. An unfortunate downside of remote work really — you can’t be sure of who’s doing the work if they don’t have to come in (and I say this as someone who’s just about to start working on a fully remote, international team). AI has made fraud massively cheaper though.

I guess at some point AI deepfakes will get so good that they look, sound, and act like a local dev while zooming and working with them throughout the day - but then I suppose what's the problem exactly?, ha.

You have to come to a certified center next to your home, where they'll check your credentials and put you in front of a certified computer with cameras recording everything.

Since they're local, they know the customs, they know which credentials to check, and they can't place an AI in front of a computer.

1) Hopping on a video call and placing the passport or (driver's license + birth cert) in front of the camera. 2) Going to an in-person business that the company is friendly with to have their HR verify you.

Both of these usually involve sending a photo of the documents in addition to the in-person/video verification.

#1 is very popular, and in a legal grey area ever since COVID. More standardized/strict businesses (banking, university) will require #2. Either one of those options kills this outright (ignoring the other billion red flags they should have seen pre-interview).

Along with a background check, which even the crappy ones will verify using things like W2s, I don't really see this as a concern, unless you've got sketchy hiring practices and hire in countries you aren't familiar with.

The candidate was very aligned to our needs and enthusiastic to work with us, so after a first call to discuss about company, motivations and so on, we scheduled another to go deeper in tech details.

I interviewed the cadidate twice, except - it wasn't the same person. When I grew suspicious and started being inquisitive, pointing out incoherences between their answers in the two interviews, the candidate said this was not a good time and basically left the call.

This is ridiculous. According to their resume they lived, studied, and worked in Poland for over 10 years and don't speak Polish? This isn't considered a red flag? This was a screening call and the head of operations categorically failed.

>Communication was a bit choppy, and the candidate did not speak Serbian, despite graduating from the University of Kragujevac, in Serbia. Once again, they sounded motivated, so proceeded to the next round.

So no actual lessons learned from the last time.

Similar to RNA interference in Cellular networks, the same thing is occurring on both parties of communications that typically happen in job hiring scenarios.

What are the incentives and who would benefit?

Certainly a nation-state following a regime-change playbook similar to the CIA would benefit from sabotage, espionage, and all that. Demoralize, Destabilize, Bring to Crisis, and Renormalize are the various stages.

Aside from that, Recruiters and professional headhunters have profit incentive to engage in these things since their services become more valuable as the noise floor increases.

This is a simple indirect imposition of additional cost on all companies seeking hiring candidates. By polluting the current candidate pool to the point of no return, the only alternative are recruiters.

The cost to do this used to be extremely high which is why it wasn't done in the past, with AI/LLMs the cost is negligible in comparison.

People today didn't want to listen about the knock-on effects of AI being generally available. Integration is happening rapidly, and the bridges have now been burnt.

Problems will only build and get worse until the instability that caused these dynamics in the first place is removed, the same general rule as any chaotic or stochastic system (such as weather/storms).

> Ask the candidate to do things that AI filters cannot handle – yet, such as:

> * A face turning side to side

With all AI ethical practices being set by unscrupulous people right now, and a gold rush of everyone else ... is an "employer" asking me to do this actually just trying to get a model to impersonate me?

Or it it an actual company, but they want training data for their AI snake oil?

Or is it a legitimate company who are being honest, but the video service they're using is going to abuse that video data (whether or not the ToU permits that, because legal isn't stopping anyone in this gold rush)?

> * standing up and walking backward from the camera, showing more body detail

> In the circumstances, It should be justifiable to run through some or all of these tests, as well as asking for all filters to be turned off.

If we stop and ask ourselves, "What would an HR expert say?"...

Well, the poor professional decorum, and setting a suspicious tone and imbalanced dynamic, the HR person might not complain about, but...

What I think think they might complain about first is us setting up the company for discrimination lawsuits and public outrage:

* What if your request comes off as sexually objectifying and/or discriminating on their basis of their physique? (Basically, "Now how about you walk back and give us a little twirl, cupcake?" or "Bro, do you even lift?")

* What if the person has mobility problems that aren't relevant to the software job?

* What if the person is trans and using filters to help them pass, due to well-known discrimination and abuse?

* What if the person routinely uses filters for a scar or a skin condition they're self-conscious about, and maybe has trauma around that? Or because they have a really unflattering camera on their personal laptop? Are they going to feel humiliated or harassed?

I'm wondering whether some of the suggestions for countermeasures might be considered OK because the scenarios they'd just talked about are of candidates in countries that can be paid a lot less for the same work. And therefore maybe thinking of them as the other or lesser, and not due the same respect we'd hopefully extend to people in our own country or another affluent country?