By John & Nisha
Whitehead
October 7,
2026
“United
States AI companies must be free to innovate without
cumbersome regulation.”—President Donald Trump,
Executive Order 14365
The era of small government is
over.
Anyone who tells you differently is selling you
a lie.
Case in point: Donald Trump
promised to shrink the federal
government.
But what if Trump’s pledge to
cut the federal work force wasn’t really about eliminating
government bureaucracy but outsourcing it to the AI tech
sector?
What if Trump, with the backing of a
self-serving industry eager to automate government, is
instead helping
to build a government that no longer needs as many human
beings to operate it?
Hundreds of thousands of
federal workers are disappearing from the government payroll
at the same time that Washington is racing to automate
the work they once performed with artificial
intelligence.
Advertisement – scroll to continue reading
Since Trump returned to office, the
federal workforce has shrunk by more than 270,000
employees, the result of hiring freezes, deferred
resignations, early retirements and reductions in
force.
At the same time, the Trump
administration has embraced an aggressive
campaign to automate government operations, with the
General Services Administration instructing agencies to
“eliminate,
optimize and automate“ their work wherever
possible.
This is not a return to smaller
government.
It is a dangerous recalibration of
government power.
What we are watching unfold is the
construction of a government that requires fewer human
beings to carry out its will, fewer human beings to question
its orders, and fewer human beings to stand in the way when
efficiency collides with legality, morality or
constitutional restraint.
That is the real danger of
automating government.
Human beings hesitate. Human
beings question. Human beings object. Human beings refuse.
Human beings blow the whistle.
Machines do not
hesitate. They do not object. They do not blow the
whistle.
For the most part, machines do what they are
programmed to do. Until they don’t.
For an
administration increasingly impatient with legal restraints,
institutional resistance and anyone who says “no,” that
is not merely a technological convenience.
It is an
authoritarian temptation.
Anyone who believes smaller
government means less government power needs to think again,
because a government run by machines is still
government.
Likewise, a government without bureaucrats
is not necessarily a government without
bureaucracy.
It may simply be bureaucracy without
human beings.
And bureaucracy without human beings may
prove far more powerful, pervasive and difficult to restrain
than anything that came before it.
What is taking
shape is a government that may employ fewer human beings
while possessing exponentially greater power to surveil,
process, classify, investigate and control the
population.
Indeed, the administration has now
launched America.gov, an
AI-powered centralized digital gateway through which
Americans will eventually interact with federal agencies and
complete government transactions. The White House boasts
that it will allow the federal government to “do
its own paperwork.”
AI can certainly reduce
backlogs, simplify forms, process information and free
workers from tedious tasks.
Certainly, artificial
intelligence can help eliminate repetitive paperwork, reduce
backlogs, identify waste, simplify complicated forms,
translate government jargon into plain English, process
enormous amounts of information and free human workers from
tedious tasks.
The concern is not whether AI can be
useful.
It can.
The danger begins when AI stops
assisting human beings and starts replacing them—not
merely as clerks but as decision-makers, investigators,
gatekeepers, analysts and enforcers of government
power.
Thus, the concern is not the technology itself
but what happens when government begins using it to
eliminate the human restraints that constitutional
government depends upon.
We are rapidly moving from a
society ruled by laws and due process to one ruled
increasingly by software.
In such a regime, the law is
no longer merely interpreted. It is executed. Automatically.
Mechanically. At machine speed.
The result is a system
in which government no longer governs. It
processes.
Your application. Your tax return. Your
disability claim. Your immigration status. Your face. Your
movements. Your associations. Your financial transactions.
Your government records request. Your threat
profile.
Perhaps eventually your guilt or
innocence.
And, if some military planners have their
way, whether you qualify as a target.
This is where
the supposed efficiency of artificial intelligence collides
head-on with the Constitution.
Government is not a
private corporation. Its purpose is not merely to become
faster, cheaper or more productive. Government possesses
powers that no corporation possesses. It can tax you. Search
you. Arrest you. Imprison you. Seize your property. Monitor
you. Restrict your liberty. And in extreme circumstances,
take your life.
And in extreme circumstances, take
your life.
When government exercises those powers,
efficiency cannot be the highest value. Liberty must
be.
Due process is inefficient. Search warrants are
inefficient. Trials are inefficient. Appeals are
inefficient. Human review is inefficient. Checks and
balances are spectacularly inefficient.
They are
supposed to be.
The Constitution deliberately places
obstacles between government power and the
individual.
Before the police may search your home,
they generally need a warrant. Before the government may
imprison you, it must give you an opportunity to challenge
the accusations. Government power is deliberately divided so
that one branch can restrain another.
All of that
takes time.
All of that requires human
beings.
All of that frustrates officials who would
prefer to simply get things done.
Yet those safeguards
are among the things standing between a free people and a
government that can act instantly, universally and without
meaningful resistance.
Artificial intelligence
promises to cut through the delays, obstacles and human
steps that slow government down.
That is
marvelous if you’re trying to schedule an appointment. It
is terrifying when the government is making decisions about
your liberty—or your life.
For generations,
government power was also constrained by manpower. There
were only so many investigators, analysts and
bureaucrats—and only so many hours in a day.
AI
changes that equation.
By doing what the police state
has historically lacked the manpower and resources to do
quickly, AI surveillance is already reorienting our world
into one in which freedom is increasingly difficult to
recognize: it can be everywhere, watch everyone and
everything, monitor, identify, catalogue, cross-check,
cross-reference and collate enormous quantities of
information.
The surveillance state no longer needs
enough people to watch everyone. It needs enough computing
power.
That distinction matters.
A human
investigator can follow only so many people. An algorithm
can analyze millions. A bureaucrat can make only so many
decisions in a day. A machine can make millions without
sleeping, hesitating or developing a conscience.
That
allows government power—and government abuse—to operate
on a scale the Founders could scarcely have
imagined.
It also creates an enormous
accountability problem.
For decades, critics of the
administrative state have warned about unelected bureaucrats
wielding excessive government power.
Replacing
the bureaucrat with an algorithm does not solve that
problem. It may make it worse.
At least the bureaucrat
has a name. A human being can be questioned. A supervisor
can At least the bureaucrat has a name. A human being can be
questioned. A supervisor can intervene. An employee can
refuse an unlawful order. A whistleblower can expose
misconduct. Someone can be held responsible.
But what
happens when government responds, in effect, that the
computer made the decision? Who do you cross-examine? Who
explains the reasoning? Who takes responsibility for the
mistake? Who possesses the authority to override the
machine? And what happens when the algorithm itself belongs
not to the government but to a private corporation whose
code, training data and internal workings are shielded from
public scrutiny?
The government’s dependence on
private technology companies only compounds the
problem.
Government power remains government power,
but accountability becomes privatized.
We are already
seeing glimpses of this future.
Federal agencies are
already experimenting
with AI to review Freedom of Information Act requests
and recommend which portions of government records should be
withheld from the public. Customs and Border Protection is
testing
AI-assisted redaction tools, while other
agencies are developing similar systems. Human reviewers
are still involved, but the trajectory is unmistakable:
tasks requiring judgment about what the public may know are
increasingly becoming candidates for automation.
The
stakes become far greater once AI moves from paperwork to
policing and warfare.
Earlier this year, the
Pentagon clashed with Anthropic, the company behind the
Claude AI system, after Anthropic insisted on maintaining
restrictions against two uses of its technology: mass
domestic surveillance and fully autonomous
weapons.
Anthropic warned that present AI systems are
not reliable enough to independently decide when weapons
should fire and argued that AI-powered mass surveillance
could allow government to piece together enormous quantities
of otherwise scattered information about Americans at
unprecedented scale.
The Pentagon
responded by designating Anthropic a supply-chain risk,
a decision that has since become the subject of
litigation.
The dispute should concern anyone who
cares about the limits of government power.
The issue
was not whether artificial intelligence could assist
military personnel.
It was whether there must always
remain a human being responsible for certain
decisions.
That question became even more urgent in
September, when the United
States and Russia opposed provisions under discussion at the
United Nations that would have required human review of
AI-generated military targets and imposed stronger
requirements concerning the predictability and reliability
of autonomous weapons systems.
The consequences of
that rush toward machine-speed warfare are no longer
hypothetical.
In February, U.S.
missiles struck the Shajarah Tayyebeh girls’ elementary
school in Minab, Iran, killing more than 150 people,
including at least 120 children.
The school had once
been part of a military compound, but it had been operating
openly as a school for years. Yet outdated
intelligence still classified the site as a military
target.
According to reporting on the Pentagon’s
internal investigation, the
flawed target was fed into Maven, the military’s
AI-assisted targeting system. Officials reportedly relied
too heavily on the technology to identify stale or
contradictory intelligence while an accelerated targeting
process compressed work that once took hours into minutes.
At the same time, civilian-harm mitigation teams had been
dramatically reduced.
The machine did not simply wake
up one morning and decide to bomb a school. Human beings
built a system that placed a premium on speed, automation
and scale—and then gave it priority over human judgment
and careful vetting.
That distinction matters. It is
also precisely the danger.
Think about the
progression: First AI does the paperwork. Then AI recommends
the decision. Then AI makes the decision.
Eventually
AI decides who lives and who dies.
At each step,
humans remain tempted to surrender a little more judgment to
the machine because it is faster, cheaper, more efficient
and supposedly more objective.
Yet technology does not
eliminate human bias, political agendas or government
misconduct.
It can automate them. It can amplify them.
It can conceal them behind a technological black box. And it
can reproduce them millions of times before anyone realizes
what has happened.
This is why assurances that AI
companies can largely police themselves should provide
little comfort.
At a White House gathering with
technology executives on September 29, Trump praised what he
called “tremendous self-policing” by the AI industry and
joined major technology companies in announcing a voluntary
safety accord emphasizing internal controls, outside
audits and corporate oversight.
In other words, Trump
wants us to trust the government and the AI sector to police
itself.
The Founders knew better than to trust
power to police itself. That is one of the central premises
of the Constitution.
As James Madison famously
observed in Federalist No. 51, “If
men were angels, no government would be
necessary.”
The Founders did not assume that
those entrusted with power would always restrain themselves.
They divided power among competing branches because they
understood that good intentions are not a safeguard against
tyranny. Neither are voluntary promises. Neither are
corporate ethics policies. Neither are
algorithms.
Machines are not angels. The corporations
building them are not angels. The politicians deploying them
are not angels. And artificial intelligence does not repeal
the oldest lesson in the history of government: power
requires limits because power will always seek to
expand.
This is where the campaign to shrink the
federal workforce takes on an entirely different
significance.
There is nothing inherently wrong with
reducing unnecessary government employment, eliminating
waste or using technology to make public services work
better.
The danger lies in confusing fewer government
employees with less government.
A government can have
fewer workers and still become vastly more powerful. It can
automate surveillance. Automate investigations. Automate
eligibility decisions. Automate regulatory enforcement.
Automate censorship. Automate watchlists. Automate
targeting. Automate punishment.
Indeed, a government
freed from the practical limits imposed by manpower could
become more pervasive than anything an earlier generation of
bureaucrats could have achieved.
That is the paradox
confronting us.
We may be building a smaller federal
workforce and a larger government machine.
Trump will
not be president forever. Neither will the politicians who
come after him.
But the AI infrastructure being
erected now—the databases, automated systems, surveillance
capabilities, digital gateways and government relationships
with technology companies—will be inherited by every
future administration.
As I make clear in my book Battlefield
America: The War on the American People and in its
fictional counterpart The
Erik Blair Diaries, once those powers become
embedded in the machinery of government, they will be
exceedingly difficult to dismantle.
That is why the
debate over artificial intelligence cannot be reduced to
whether AI creates jobs, destroys jobs, saves taxpayer money
or makes government more convenient.
The greater
question is what happens when fewer and fewer human beings
stand between government power and the people subjected to
it.
What happens when there is no human face behind
the decision? No human judgment. No human hesitation. No
human mercy. No human conscience.
No human being
willing—or able—to say: this is wrong.
The great
danger is not that artificial intelligence will suddenly
seize control of the government. It is that government
officials will willingly surrender more and more control to
artificial intelligence because machines are faster,
cheaper, more obedient and easier to blame.
That is
how bureaucracy without bureaucrats becomes authoritarianism
without accountability.
The government may employ
fewer people. It may operate fewer offices. It may eliminate
paperwork and automate entire agencies. But if it can watch
more people, investigate more people, process more people,
target more people and exercise more power over more lives
than ever before, that is not smaller government.
It
is government without human limits.
The Founders built
the Constitution around a simple truth: power cannot be
trusted to restrain itself.
That truth does not become
obsolete because the machinery of power now runs on
artificial intelligence. If anything, it becomes more
urgent.
Because once we remove the human beings who
can question, resist, object and refuse, we may discover too
late that we have not eliminated the bureaucracy at
all.
We have eliminated the
brakes.
Constitutional attorney and author John W.
Whitehead is founder and president of The Rutherford
Institute. His latest books The
Erik Blair Diaries and Battlefield
America: The War on the American People are
available at www.amazon.com. Whitehead can be contacted at
johnw@rutherford.org. Nisha Whitehead is the Executive
Director of The Rutherford Institute. Information about The
Rutherford Institute is available at www.rutherford.org.
