Imagine one day,
your manager walks into a meeting and tells you:
"AI can do your job now."
That’s already a nerve-wracking statement.
But if the next sentence is:
"And AI is cheaper than you,"
then the issue changes completely.
Because now the question isn’t:
Is AI really advanced enough?
Instead, it’s:
Can a company simply replace a person whenever it finds a cheaper technology?
A recent labor dispute case in Hangzhou, China, brought this question directly into the courtroom.
What’s most notable is that the affected position wasn’t a role unrelated to technology at all.
The person was originally:
Someone working for AI.
His job was to check AI for mistakes
In the case, Mr. Zhou, 35 years old,
served as a quality control supervisor for an AI large language model at a technology company in Hangzhou.
Simply put,
when users ask questions to a large language model,
the model’s answers cannot just be released without review.
Someone needs to check:
Are the answers correct?
Are there any violations?
Are there privacy issues?
Which type of model or handling is appropriate for different questions?
Mr. Zhou’s role was essentially:
AI quality control.
His monthly salary was:
25,000 RMB.
But as AI technology advanced,
the company began to believe
some of the quality control work done manually
could now be handled by AI itself.
This led to a very AI-era scenario:
The person checking AI started being replaced by AI.
The company didn’t immediately fire him but first proposed a demotion and pay cut
In 2025,
the company first approached Mr. Zhou with a transfer proposal.
From a supervisor role,
they wanted to move him to a more general operations job.
The biggest issue was the salary.
Originally,
25,000 RMB monthly,
the new offer was:
15,000 RMB monthly.
This was a
40% pay cut.
Mr. Zhou refused.
They couldn’t reach an agreement.
The company then notified him of
termination of the labor contract.
One core reason the company explained later was:
With AI technology upgrades,
the quality control tasks he handled could be automated
and AI also had a cost advantage.
From a business perspective,
this logic is easy to understand.
If a task once required paying a person 25,000 RMB monthly,
but AI can now do most of it at a lower cost,
of course the company would reassess
whether that position is still necessary.
But just because it makes sense financially,
does not mean it’s legally permissible to proceed this way.
He filed for labor arbitration and won through the second court hearing
Mr. Zhou did not accept the company’s handling.
He applied for labor arbitration,
demanding responsibility for the illegal termination.
The arbitration sided with him.
The company disagreed and sued in court.
In the first trial,
the company lost again.
They appealed.
In the second trial,
the outcome remained unchanged.
Ultimately, the court ruled:
The company unlawfully terminated the labor contract,
and must pay
over 260,000 RMB in compensation.
The crucial point here is not the 260,000 amount,
but the court addressing a question that may become more frequent:
Is "AI being cheaper" alone sufficient reason to dismiss employees?
The answer in this case was:
No.
The company tried to justify termination citing "significant changes in objective circumstances"
Article 40 of China’s Labor Contract Law provides for terminating a labor contract
when the "objective circumstances" underpinning the contract
have significantly changed
making it impossible to fulfill the original contract,
and after negotiation, no new agreement is reached.
This can apply to situations like
company relocations,
mergers,
or major asset transfers,
which are external, uncontrollable changes.
The legal question in this case becomes:
Is the company’s decision to implement AI considered a "significant change in objective circumstances"?
The Hangzhou court’s answer:
No, not so easily.
Because AI didn’t just fall from the sky
This is the key takeaway from the case.
Using AI is a move to
improve efficiency,
reduce costs,
and boost competitiveness.
This is a fully reasonable business decision.
But the court sees such technology upgrades as
an active choice by the company’s management,
not a sudden, uncontrollable event like
earthquakes,
force majeure,
or forced relocation.
In other words,
a company cannot just decide:
"We want to use AI to cut costs,"
then claim:
"External conditions changed suddenly, so your contract can’t continue."
These are fundamentally different.
More importantly: Does the original job really cease to exist completely?
The court also considered whether, even if AI can automate some tasks,
the company demonstrated that
Mr. Zhou’s original contract became entirely impossible to fulfill.
The court rejected the company’s claim.
Furthermore, the company’s alternative offer,
cutting monthly salary from 25,000 to 15,000 RMB,
was deemed an unreasonable negotiation proposal.
This case isn’t about
an employee simply refusing AI and the court stepping in to protect them.
Instead, it shows that when technology changes job roles,
the company must prove more, such as:
How much has the original position changed?
Which tasks still require humans?
Are there other reasonable positions available?
Can training help?
Is the reassignment fair?
Have both sides genuinely negotiated?
All these can’t be replaced by a simple statement:
"AI can do it now."
The court even shifted more responsibility onto the company
When the Hangzhou court released this case,
it pointed out an important direction.
Companies can and should use AI to
improve efficiency,
transform,
and redesign jobs.
But the costs caused by technological progress
cannot be completely borne by employees by default.
If job duties truly change,
companies should consider first:
training,
skills upgrading,
job reassignments,
and respecialization,
allowing existing employees
the chance to transition into new roles.
Not simply:
"AI got better, so you’re out."
This logic is highly important for all companies.
Because when AI fully enters workplaces,
it’s almost certain that
jobs may disappear, but it doesn’t mean all employees should.
The more common scenario is not "the whole occupation disappears overnight" but a job being gradually divided
Before, many discussions asked:
Will AI replace accountants?
Will AI replace engineers?
Will AI replace analysts?
That’s too broad a question.
In reality,
it’s rarely the case that
an occupation exists on Monday,
and disappears entirely by Tuesday.
What often happens is:
A position involving ten tasks,
AI takes over four first,
then two more,
until the company asks:
Do the remaining four tasks require a full position?
This is the real challenge many white-collar jobs face today.
A Beijing research analyst faced a different version of the impact
South China Morning Post reported on September 4 the experience of Clare Zhang,
a research analyst in Beijing.
She conducted corporate research for four years,
including:
reviewing company filings,
looking up public records,
analyzing markets,
studying competitors,
and finally producing comprehensive reports,
sometimes up to
500 pages long.
This work used to require extensive manpower,
because the time was mostly spent on
collecting data,
categorizing,
organizing,
cross-checking,
creating charts,
and writing background information.
But these areas are exactly where generative AI has rapidly improved.
AI can now more quickly:
gather data,
organize information,
analyze numbers,
and generate visuals.
Zhang said when her company later laid her off,
HR explicitly mentioned AI,
stating AI made it unnecessary and financially unreasonable to keep so many analysts.
Two other colleagues were laid off afterwards.
This shows a very practical distinction ahead.
Courts can regulate how layoffs occur, but can’t guarantee jobs always exist
The Hangzhou case can be easily misunderstood as:
"Chinese courts banned AI from replacing employees."
That’s not true.
The court didn’t say
companies can never adjust their workforce when implementing AI.
The real judgment was:
In this case, the company can’t just say
"We introduced AI; it’s cheaper, so this counts as a significant change in objective circumstances,"
and use that as a reason to cut salary and terminate contracts.
China’s Labor Contract Law also has provisions for economic layoffs and major technological or business adjustments,
but these require their own legal conditions and procedures.
So the true message is not:
"The law can stop AI."
But rather:
"When technology changes jobs, companies must still follow rules on handling affected employees."
For employees, the real risk may be slower than an immediate AI-triggered dismissal
The biggest impact of AI on jobs
is not always a sudden:
manager telling you,
"Don’t come in tomorrow."
It could be a series of gradual changes.
A team of ten people,
after AI is introduced,
shrinks to eight,
then six.
The company may avoid mass layoffs,
simply not replacing people who leave,
hiring fewer newcomers,
stopping outsourcing,
and reducing junior positions.
Work once done by three people may become done by one person plus AI.
This kind of change
rarely makes legal news,
but its impact on the labor market
may be even greater.
Even companies that "do work for others" are being squeezed by AI
This is very similar to consulting, a topic SasaDaily discussed recently.
Companies used to outsource
IT maintenance,
research,
software development,
testing,
data analysis
to external firms.
Those firms’ business model relied on
many people,
charging for many hours.
But as AI Agents shorten part of the work,
clients start to question:
If AI can do this,
why pay so many hours for so many people?
Ultimately, it may not be
an individual employee first laid off,
but
the entire contract disappears.
Zhang also mentioned that some clients of her former company
brought the research work back in-house,
pairing with AI to do it themselves.
At that point, AI’s impact
was not just on one employee,
but on the entire chain:
client → consulting firm → research team → junior analyst.
That’s why just "learning to use AI" may not be enough
Many people hear:
AI will change jobs,
and their immediate reaction is:
I better learn ChatGPT quickly.
That’s not wrong.
But if your entire value is
turning a four-hour task
into a one-hour task using AI,
the company’s next question may be:
Then why do you need three people?
So the real upgrade isn’t just
knowing how to operate AI,
but understanding:
Which tasks can AI handle,
and which still require your judgment?
Who is responsible for verifying AI’s output?
When exceptions happen, who knows where the problem lies?
What does the client really want to solve?
When AI offers five solutions,
who decides which to use?
These abilities
may determine how much value remains in a role after the job is redefined.
For companies, AI implementation shouldn’t instantly translate to fewer people
This case also serves as a reminder to employers.
Introducing AI
might prompt simple Excel calculations like:
10 people now,
AI improves efficiency by 30%,
so maybe 3 fewer employees?
But real work doesn’t always work that way.
If AI takes over tasks like
organizing,
searching,
categorizing,
and drafting first versions,
the remaining staff must handle more:
judgment,
exception handling,
quality assurance,
client communication,
and risk management.
New roles also emerge:
Who verifies AI results?
Who manages access rights?
Can data be given to AI?
Who’s responsible for errors?
Who retests after model updates?
This is why mature AI transformation
isn’t about firing someone as soon as the AI does a step,
but first asking:
After redesigning the work, where should people fit in?
This dispute isn’t really "Human vs AI"
Can AI take over many jobs?
The answer is probably:
Yes.
And even more in the future.
Will companies recalculate staffing for this reason?
Absolutely.
Law can’t limit technological progress itself.
What law can regulate is:
how companies handle existing employees affected by technological changes.
The boundary drawn by the Hangzhou case is clear:
AI being cheaper is a business reason,
but a business reason
does not automatically become a legal reason for dismissal.
Technology can rewrite work.
Companies can redesign their organizations.
But at least in this case,
the court ruled:
a company cannot unilaterally decide to cut costs with AI
and then transfer all the risks of technological upgrades
onto employees.
If one day your company really tells you:
"AI can do your job now,"
the real question might not just be:
Am I about to be replaced?
But rather:
Which tasks still need humans, and will the company give me a chance to move into those roles?
If it were you,
after your company finds AI can do most of your work,
would your first step be to
lay people off,
or
train, reorganize, and reasonably reassign?
Today, grow a little with AI.
Learn one AI skill every day.
Save a little time every day.
Improve a bit every day.
SasaDaily, growing with you.
Recommended Reading
AI saves your work time, so why do bosses want you to do more?
