In the past, the convenience of remote interviews was:
No need to meet in person.
The company might be in Taipei,
and you could be in Tainan.
Or the company could be in the US,
while you were in Asia.
Just turn on your camera,
chat for half an hour,
and the company could start judging:
Does this person have the skills for this job?
But by 2026,
some remote interviews suddenly include previously unusual requests.
"Could you turn off your virtual background?"
"Can you pan your camera around so I can see the surroundings?"
Even:
"Can you wave your hand in front of your face?"
These aren’t interview questions.
The company isn’t suddenly interested in your room.
What they really want to know is:
Is the person in front of the screen actually present and truly who they claim to be?
The New First Step in Remote Interviews: From Testing Ability to Verifying Identity
On August 31, the Wall Street Journal reported
that as AI-assisted answers, deepfake technology, identity fraud, and remote job scams increase,
some companies are strengthening the verification process for remote hiring.
Interviewers have started asking candidates to:
remove their virtual backgrounds,
move their cameras,
wave their hands in front of their faces,
and even have their login locations and IP addresses checked.
Such requests once seemed absurd.
During a normal interview,
companies cared about:
what you’ve done,
how much you know,
and whether you can solve problems.
But now an extra layer has been added to the hiring process:
"Are you really who you say you are?"
Why has waving your hand across your face become part of the interview?
Because real-time face-swapping or deepfake systems
need to continuously map a generated or replaced face
onto a real person’s image.
If the scene is simple,
the person sits facing the camera,
the lighting is fixed,
the system can look very natural.
But if:
a hand suddenly covers part of the face,
the head moves quickly,
lighting changes,
or the person moves,
the system must process in real time:
Which part is a hand?
Which part is the face?
What needs to be hidden?
What should reappear?
So some recruiters have started using these simple actions
to look for unnatural edges,
delays,
distortions,
or suspicious glitches in the video.
This is not a perfect "real-person test."
But it shows one thing clearly:
Companies can no longer assume "a face on screen" means "the person in front of the camera is genuine."
What's even more troubling: some fake candidates pass technical interviews
If AI just helps answer a few questions,
that’s already complicated enough.
But Anastasios Angelopoulos, CEO of AI model evaluation firm Arena, publicly shared in August,
they’ve faced even more extreme cases.
Some applicants:
attend interviews,
answer technical questions,
communicate with engineers,
and even pass technical interviews.
The engineers genuinely believe
they are speaking to a real candidate.
But upon preparing to hire,
the company realizes something is wrong.
Angelopoulos described
that when they intended to bring the person on board,
the candidate never actually appeared in person.
He even stated:
"This person wasn’t real—it was AI."
This sounds like science fiction.
But importantly,
Arena hasn’t publicly explained how they confirmed these candidates were AI-generated.
So this shouldn't be interpreted as:
"Now many fully AI-generated virtual candidates are landing jobs."
A more accurate description is:
Companies have encountered fake candidates convincing enough to fool interviewers, only discovering serious identity issues afterward.
So Arena is now considering a very traditional method: meeting in person for a handshake
After encountering AI-generated fake candidates,
the solution isn’t:
to buy an even more advanced AI.
It is:
to come into the office.
Angelopoulos said Arena is now considering requiring new hires to:
physically visit the office
if they want to receive company equipment.
Meet face-to-face.
Shake hands.
Confirm the person truly exists.
What’s interesting here is that
for more than a decade,
technology has been proving:
many things can be done without meeting in person.
Online banking.
Remote healthcare.
Video meetings.
Online contracts.
Remote work.
Remote interviews.
But as AI makes "remote presence" more realistic,
some companies are rediscovering:
Seeing someone in person again has become a valuable way to verify identity.
This isn’t just companies overthinking—FBI officially warns about identity risks
The riskiest cases even involve national security.
The US FBI issued a warning in 2025,
specifically urging companies to strengthen their remote hiring processes.
The alert mentioned:
North Korean IT workers have used
AI and face-swapping technology
in video interviews to conceal their real identities.
Therefore, the FBI recommends companies:
verify identities during interviews,
verify again at onboarding,
and continue identity checks even after remote work starts.
The FBI also advises:
cross-check resumes and contact information,
look out for the same phone numbers or emails used by different applicants,
ask detailed questions about location and education background,
and try to conduct parts of hiring and onboarding processes in person when possible.
This means that:
"Verifying who is in front of the screen" is no longer just an HR issue.
It may also be:
a cybersecurity problem,
a company data issue,
and even a matter of national security.
Because fake employees might want more than just a paycheck
Suppose someone only wants to use AI to answer interview questions
to get a job they are actually unqualified for.
That’s already hiring fraud.
But a more dangerous scenario is:
The real goal is to gain access to:
company accounts,
VPNs,
GitHub repositories,
source code,
internal documents,
customer data,
and company computers.
Once someone is hired,
the company often proactively grants these permissions.
That means attackers might not need to
"hack into" the company.
Instead, they can:
"get hired by the company."
This is why identity verification has suddenly become so important.
The hiring process itself
may be the entry point into the company’s internal systems.
AI makes "looking like the real person" increasingly cheap
In the past, if you wanted to impersonate someone in a video interview,
it was very difficult.
You’d need to find someone who looks quite similar,
or have someone else attend for you.
Your voice would likely be different too.
Ask deeper questions about the resume,
and you’d likely get caught.
But generative AI slowly combines many previously separate technologies.
Now it can handle:
real-time voice generation,
voice imitation,
face swapping,
background generation,
real-time interview question prompts,
coding assistance,
and resume creation.
Each capability alone may have legitimate use.
But combined,
the problem faced by recruiters is very different.
Whereas interviewers once saw simply:
a person,
now they might be facing:
a person + AI + real-time assistance tools + synthetic identity
presenting results together.
However, an important boundary remains: Using AI isn’t inherently cheating
This issue must not swing to the opposite extreme:
"Anyone using AI in an interview is fake."
Because AI has become a normal work tool.
Some people use AI to:
organize resumes,
practice interviews,
improve English,
prepare questions,
and organize company info.
Some companies even allow candidates to use AI for
take-home coding tasks
or certain practical work simulations.
Because what companies really want to assess is:
how you will perform the job after joining.
If AI use is permitted during work,
then outright banning AI for candidates during all stages
may not be reasonable.
The real distinctions to draw are between:
AI helping you prepare,
AI helping with your work,
and
AI impersonating another person.
The first two can have rules and discussions.
The third involves issues of identity itself.
Hiring processes might split into two tests in the future
Future remote hiring may no longer only ask:
"Do you have the skills?"
It will first address another question:
"Who are you?"
The first layer:
identity verification,
confirming:
the actual person,
location,
basic info,
equipment, and account consistency.
The second layer:
skill verification,
checking:
your abilities,
how you think,
how much AI you use,
and what you understand without AI.
These two have often been mixed together
because we assumed:
the person answering was the person on the resume.
But as that assumption becomes unreliable,
the entire hiring process must be redesigned.
This also explains why in-person problem-solving is gaining value again
Companies once preferred take-home assignments.
Complete a presentation at home.
Write a piece of code.
Finish an analysis.
Submit it days later.
The advantage was:
candidates had time to think thoroughly,
and companies could see work closer to real tasks.
But now AI can generate answers,
fix code,
write analyses,
and create presentations within seconds.
This doesn’t mean take-home assignments are useless.
They just can’t alone answer:
"How much of this work did the candidate actually do?"
So some companies are again emphasizing:
on-the-spot problem-solving,
real-time questioning,
explaining thought processes,
and face-to-face conversations.
Not because AI is necessarily prohibited,
but because:
companies need to know what remains when some tools are removed.
Candidates may also face new costs
This situation poses a risk to companies,
but isn’t without cost for honest candidates either.
You may increasingly be asked to:
turn on your camera,
disable backgrounds,
verify your location,
present identification,
solve problems on the spot,
or even appear at the office for identity confirmation.
Companies will say:
this helps prevent fraud.
But candidates have the right to ask:
what data is collected?
How long is it stored?
Who can access it?
Which service handles identity verification?
Is there a risk of misidentifying genuine people as suspicious?
Interestingly, this shares a common theme with recent stories about AI detectors:
When technology helps determine "real or fake," those being judged should also have opportunities to explain and appeal.
Ultimately, what may become rare isn’t answers but a trustworthy identity
One aspect rapidly becoming cheap in the AI era:
Generating answers.
Resumes can be generated.
Emails can be generated.
Code can be generated.
Voices can be generated.
Faces can be generated.
Even:
"someone who looks like they’re interviewing you"
can be partially simulated.
When "looking real" becomes easier,
another factor becomes more valuable:
Verified authenticity.
You truly are who you say you are.
Your experiences really belong to you.
This work was genuinely done by you.
This judgment was truly made based on your understanding.
So the biggest change AI brings to hiring may not be:
companies using AI to select candidates,
or job seekers using AI to find jobs.
More deeply:
what we never had to prove before—"I am me"—now needs to be verified.
The greatest breakthrough of remote interviews used to be:
No meeting required, yet you can trust the other party.
AI’s new challenge is:
After seeing, hearing, and talking to the person for half an hour, can you still be certain it’s really them?
Today, grow alongside AI.
Learn one AI skill each day.
Save a bit of time every day.
Improve your ability a little more each day.
SasaDaily, growing with you.
Recommended Reading
Why Does AI Often Score Resumes Before Humans Even See Them?
Why Did AI Detectors Mark a Short Story as "100% AI" Even After Review?
