If the Future Is for Everyone, Who Protects Everyone?

An Open Response to Mark Zuckerberg on Freedom, Jobs, Power and the Guardrails of Superintelligence

By Tiffany Johnson-Wilson

At 2:00 in the morning, I found myself sitting in front of a screen, using artificial intelligence to build a real estate technology company, while reading Mark Zuckerberg tell me that superintelligence will improve our quality of life.

The irony was impossible to miss.

I am probably not the person anyone would expect to write a rebuttal to an optimistic vision of artificial intelligence. I am a small business owner. I am building technology that I believe can fundamentally change the way an entire industry operates. AI has allowed me to learn things I did not know, build things I could not have built as quickly before, challenge established assumptions in my business, and compete in ways that would once have required far more people and far more capital.

In many ways, I am living the future Zuckerberg describes.

And I am excited about it.

That is exactly why I think those of us building that future have an obligation to question it.

In his August 10 essay, The Future Is for Everyone: The Path to a Positive AI Future, Zuckerberg proposes a philosophy for superintelligence built around three principles: individual empowerment as a source of prosperity, invention as the primary purpose of superintelligence, and a balance of power as the foundation of safety.

There is much in that philosophy that I agree with.

But there is one question I kept writing in the margins as I read:

How far does the power go?

And eventually another:

Who protects everyone from the power we are giving everyone?

Will AI improve our quality of life or merely optimize it?

Zuckerberg begins his essay with an extraordinary promise. Superintelligence, he predicts, will allow people to create businesses, discover new things, express ideas, learn, improve health and improve our quality of life. Later, he imagines personal agents helping with relationships, health, careers, finances, households and hobbies while devices like AI glasses allow us to remain present with the people around us.

I understand the argument about health. I understand the argument about productivity. I certainly understand the argument about business creation.

But I paused at “quality of life.”  That made me tilt my head alittle. 

What exactly does he mean by quality of life in a world saturated with artificial intelligence?

If our agent is thinking about our work, our health, our money, our sleep, our exercise, our relationships, our hobbies and our next decision, have we freed ourselves from technology – or simply became more robotic under technology we can no longer see?

Maybe removing the phone from our hand and putting intelligence into glasses really will allow us to be more present.

Or maybe it will mean the computer never leaves the room at all.

A more efficient life is not automatically a better life.

A more optimized person is not automatically a happier person.

And a world in which every business, employee, student and entrepreneur can accomplish exponentially more may not result in exponentially more leisure. It may simply change the baseline of what everyone is expected to accomplish.

I know because I can already feel it.

I can use AI to do things at midnight that once would have required an employee, a consultant or an entire department. That is extraordinary.

It also means the entrepreneur next to me can do the same thing.

So I work later.

If everyone becomes ten times more productive, does everyone suddenly get Friday afternoon off?

Or does ten times the output eventually become the expectation?

That distinction matters if we are going to promise people a better quality of life.

Empowerment is a value. It is not, by itself, a safety system.

Zuckerberg is deeply skeptical of concentrating superintelligence in the hands of a small group of experts, corporations or governments.

On that point, I agree with him.

History gives us plenty of reason to be suspicious of concentrated power. I believe in free enterprise, individual liberty, open inquiry and the right of ordinary people to pursue extraordinary ideas.

I would not be building what I am building if I did not.

But distributing power does not make power harmless.

The power to invent is not inherently good or bad.

One person may use extraordinary intelligence to create a treatment that saves a child with cancer. That promise is personal to me: I lost my brother to a brain tumor, and if superintelligence could have found a cure we did not have, I would have wanted every ounce of intelligence we could bring to the problem.

Another may use similar capabilities to explore how a dangerous biological agent could be created.

Zuckerberg himself acknowledges this dual-use problem when he discusses biological and chemical risks.

Because Zuckerberg frames his vision for superintelligence as a philosophy, for me the philosophical question becomes more complicated than simply whether intelligence should be centralized or distributed.

The issue is not that I have been given a powerful tool to shape my future.

The issue is that I may have been given a powerful tool capable of affecting your future.

Those are different categories of freedom.

My freedom to build a business is one thing.

My freedom to take an action capable of producing catastrophic consequences for people who never consented to that risk is another.

Every functioning free society recognizes this distinction.

Free enterprise does not mean a restaurant can ignore food-safety laws.

Individual liberty does not eliminate aviation rules, building codes, securities laws or criminal statutes.

We establish boundaries around powerful things precisely because we value freedom enough to protect the society in which freedom exists.

Zuckerberg does not actually advocate a world without boundaries. He explicitly recognizes legal and safety limits, government policy and independent oversight.

My disagreement is subtler.

I am not convinced that broad distribution plus competing interests equals an adequate system of checks and balances.

Competition is not the same thing as checks and balances.

This may be the part of Zuckerberg’s argument I find most intellectually interesting.

He asks us to imagine one person with a superintelligent lawyer. That person would possess an extraordinary advantage. Give everyone a superintelligent lawyer, he argues, and the imbalance disappears.

He applies similar reasoning to cybersecurity and business: broadly distribute superintelligence and competing parties will check one another.

There is logic to this. But democratic checks and balances are not created merely by giving competing people equal amounts of power.

We have constitutions.

We have statutes.

We have courts.

We have due process.

We have enforcement.

We have institutional boundaries determining what even powerful people are permitted to do.

Competition is one component of a free society. It is not a replacement for governance.

Two enormously powerful actors can check each other.

They can also escalate each other.

That distinction becomes particularly important when Zuckerberg later describes AI as perhaps the most competitive industry in history and warns that advantages can be measured in months—or even a single month.

Competition may protect us from monopoly.

Competition may also be the thing pushing everyone to run faster.

And if the technology that is supposed to improve our quality of life creates a world where none of us feels permitted to slow down, then I think we have to ask what we mean by “quality of life” in the first place.

Zuckerberg may answer one of his own questions

Near the beginning of his essay, Zuckerberg asks why anyone who truly believes AI could eliminate most jobs or diminish humanity’s relevance would nevertheless rush to build that future.

I wonder the same thing. Then, much later in the essay, I thought he partly answered his own question.

When discussing recursive self-improvement, Zuckerberg explains that once AI systems can substantially improve themselves, a laboratory that refuses to allocate enough computing capacity toward self-improvement risks falling behind competitors.

There it is.

Why do people race toward technology whose consequences they themselves admit may be profound?

Because someone else is racing too.

Because nobody wants to be second.

Because companies want to win.

Countries want to win.

Founders want to win.

Investors want returns.

Researchers want breakthroughs.

Human beings want recognition, prosperity, security, accomplishment and sometimes simply the satisfaction of being first.

Zuckerberg writes elsewhere that empowered people naturally compete with one another.

Exactly.

That competitive instinct may be one of the forces accelerating AI development in the first place.

And that raises a difficult question:

What happens when the mechanism Zuckerberg relies on to keep power balanced is also the mechanism forcing development to move faster?

The “natural balance” of jobs deserves more scrutiny.

Zuckerberg argues that there is a natural economic balance between businesses using automation to serve people’s needs more efficiently and individuals gaining capabilities to meet more advanced needs.

I want that to be true.

But is it actually a natural law?

The available evidence is more complicated.

The World Economic Forum’s 2025 Future of Jobs Report, which considers AI alongside other major economic forces, projected 170 million jobs could be created by 2030 while 92 million could be displaced, a net gain of 78 million.

That sounds encouraging.

Until you look at the transition.

The same report estimated that, out of every 100 workers, 59 may require training by 2030 and 11 may not receive the retraining or upskilling they need. It also reported that 41% of surveyed employers expected to reduce their workforce where AI can automate tasks.

The International Labour Organization offers another reason not to descend into doom: its 2025 work found that transformation, rather than outright replacement, remains the more likely outcome for many jobs.

But it also estimated that one in four workers worldwide is in an occupation with at least some exposure to generative AI.

None of those statistics proves Zuckerberg wrong.

They do prove something else:

A positive net employment number is not a transition plan.

If 170 million jobs are created and 92 million disappear, the person whose job disappears on Tuesday is not automatically qualified for one of the new jobs on Wednesday.

The new job may be in another city.

It may require different education.

It may pay differently.

It may require technological skills the displaced worker does not possess or does not want to spend the next several years acquiring.

The one-person-company paradox

Zuckerberg gives examples of future careers that sound incredible.

One-person product studios designing customized furniture, toys or clothing.

Experience designers creating entirely new worlds.

Personal biologists developing individualized therapies.

I want some of this future.

I would gladly volunteer for the personal biologist trying to figure out how to make me age backward.

But look closely at the phrase one-person studio.

It is impressive because that same work once required much more than one person.

That is the innovation.

It is also the labor displacement.

Zuckerberg predicts that company sizes may shrink while the number of companies increases, producing more total employment and allowing small groups of people to operate enterprises at substantial scale.

Perhaps he will be right. As a small business owner, I would love to see an explosion of entrepreneurship.

But the equation is not complete.

If an old business needed 100 employees and an AI-enabled business needs 10, we need enough additional businesses and sustainable ones to absorb the difference.

More companies with fewer employees could mean more jobs.

It could also mean dramatically more economic output produced by dramatically fewer people.

Those are not the same future.

Entrepreneurship, productivity and employment are related.

They are not synonyms.

I know this contradiction personally because I am building technology I believe will transform the entire real estate industry.

I believe excellent real estate agents will become dramatically better with AI.

I also believe some work currently performed by average real estate agents will no longer need to be performed by average real estate agents.

It would be intellectually dishonest for me to celebrate one without acknowledging the other.

Maybe some job destruction is simply part of technological progress.

Human civilization has survived enormous occupational transitions before.

But if we believe displacement is foreseeable, then those of us benefiting from this technology have a responsibility to discuss the consequences before they arrive—not after.

And that brings me to a group that appears strangely absent from much of the conversation about our AI future:

the average person.

What happens to people who do not want to become extraordinary?

Our culture celebrates adaptability.

Learn another skill.

Start another company.

Reinvent yourself.

Become more efficient.

Use AI.

Keep up.

I believe in all of those things.

But what happens to the person who does not want to continuously reinvent himself?

What about a 63-year-old worker who wants to work comfortably for another five years doing something he already knows how to do?

What about someone who is simply not interested in artificial intelligence?

What about the person who is perfectly content earning an ordinary living, raising a family, coaching a baseball team, reading a book on Saturday and never creating a new invention or process or founding a company?

Does everyone have an obligation to maximize his or her potential by AI standards simply because technology makes maximization possible?

Zuckerberg’s future is filled with extraordinary people using extraordinary tools to become even more capable.

But a humane economy also has to have room for ordinary people.

The people whom business culture may quickly label as “slackers” still need food, housing, purpose, community and dignity.

Not everyone wants to win the race.

Some people simply want to have a life.

And if technology changes the labor market faster than some people can or wish to change themselves, then the builders of this future cannot simply say that the next generation of jobs will eventually appear.

We have to think about what happens in between.

That means workforce retraining, certainly.

It may also mean rethinking affordable housing, wage supports, education, apprenticeships, food security and all the community services that absorb the consequences of economic dislocation.

It means anticipating pressure on mental health and addiction services rather than acting surprised if communities struggle after major employment transitions.

Innovation has a responsibility to look downstream.

Invention or acceleration?

Zuckerberg makes a distinction I like very much: the greatest contribution of superintelligence should be invention rather than automation.

That is inspiring.

Then his essay moves into biotechnology and argues that society should accelerate development of cures and update processes at the FDA and other regulators so they can keep pace with faster drug discovery.

I do not think those positions are inherently contradictory.

I do think they expose the central tension running through the entire essay.

How do we increase the speed of invention without allowing speed itself to become the governing value?

Regulatory systems can absolutely become outdated.

AI may allow us to make clinical research and regulatory review far more intelligent and efficient.

But medicine is also a field where being slower can sometimes represent deliberation rather than incompetence.

The goal should not be slow regulation.

The goal should not be fast regulation.

The goal should be good regulation that can move faster without confusing competitive urgency with evidence of safety.

That distinction becomes even more important when Zuckerberg argues that slowing American AI model releases by even a month could carry geopolitical consequences.

When a month becomes an eternity

One of the most striking themes in Zuckerberg’s essay is speed.

He argues that AI may be the most competitive industry in history and that even a small technological lead can be extraordinarily valuable. He returns to that one-month window as an example of just how consequential even a brief delay could become if foreign competitors continue advancing.

I strongly agree with him on one fundamental point:

America must remain at the forefront of artificial intelligence.

I do not want the defining technology of the next century governed primarily by authoritarian regimes.

But I also cannot ignore what this race feels like farther down the economic ladder.

At two in the morning, I am building my real estate prop tech company with AI.

Somewhere else, another company has 50 engineers working around the clock.

Another may have 500.

When the largest technology companies in the world tell us that even being one month behind can matter enormously, imagine what that mentality does to a small entrepreneur like me. 

Are these tools liberating us—giving us the better quality of life he claims?

Or are we creating an economy in which standing still for thirty days becomes professionally dangerous and, eventually, personally threatening to our financial security and quality of life?

That is not an argument to slow America indiscriminately.

It is an argument to recognize that relentless acceleration has human consequences too.

Maybe there are rare categories of frontier capability where government defenders should receive a temporary head start before unrestricted distribution.

Interestingly, Zuckerberg moves somewhat in that direction himself. He proposes that leading AI laboratories provide the government with intermediate model checkpoints and technical support so critical systems can be hardened before models are broadly released.

I think that proposal deserves more attention. Sometimes empowering the public and empowering the defender do not have to occur at precisely the same second.

Privacy is freedom. But trust is not the same thing as alignment.

I strongly agree with Zuckerberg’s defense of privacy.

A personal AI capable of understanding our finances, relationships, health, communications, ambitions and vulnerabilities could become one of the most intimate technologies human beings have ever created.

Zuckerberg says Meta plans to build a fully private mode in which even Meta cannot access a person’s information. I hope that becomes real, durable and independently verifiable.

Meta has already moved from philosophy toward product. On September 8, it launched Muse, describing it as a personal AI agent designed to act on a user’s behalf while providing controls over connected services and sensitive actions.

So this debate is no longer entirely theoretical.

Zuckerberg’s alignment theory is particularly interesting. Rather than forcing one centralized set of values onto everyone, he wants personal agents aligned with each individual’s goals and values. He argues that if billions of people trust, use and scrutinize these agents, that would create the balance of power necessary for a safer future.

This is where I remain unconvinced.

Mass adoption proves adoption. It does not necessarily prove societal alignment.

Billions of people trusting something can demonstrate usefulness.

It can demonstrate product-market fit.

It can demonstrate convenience.

It does not automatically demonstrate that the aggregate consequences are good for humanity.

Individually rational behavior can create collectively harmful outcomes.

An AI perfectly aligned to my personal objective might help me outperform a competitor.

A competitor’s perfectly aligned AI may help them outperform me.

That competition may produce innovation.

It may also create incentives neither of us would have chosen in isolation. We may begin lowering prices to beat each other, offering more and more for less, working longer hours or constantly finding new ways to outproduce one another—not because either of us independently decided that was a better way to live or run a business, but because neither of us can afford to stop while the other keeps going.

Eventually, the competition that makes us more productive can also make both of us less profitable, more consumed by work and further from the quality of life the technology was supposed to improve.

Personal alignment therefore answers one question:

Does my agent work for me?

It does not fully answer another:

What happens when what works for me imposes costs on everyone else?

Privacy deserves the highest level of protection.

But privacy should not become synonymous with impunity.

I do not want a surveillance state monitoring every conversation with artificial intelligence.

I also do not believe the only choices available to us are total surveillance or zero enforceable limits.

Free societies have spent centuries developing concepts such as warrants, due process, probable cause, judicial review and narrowly tailored government authority.

Superintelligence may require us to modernize those principles.

It should not require us to abandon them.

Data centers make the philosophical debate physical

I want to commend Zuckerberg and Meta for something else.

The Future Is for Everyone Fund and America’s Workforce Academy recognize that companies building infrastructure at this scale have obligations to the communities hosting it. Zuckerberg argues that data-center communities should receive high-paying jobs, public investment, infrastructure benefits and environmental commitments.

That is the right conversation. With great power comes great responsibility, and the amount of power held by the companies developing frontier artificial intelligence is vastly greater than the power held by a business like mine. Responsibility should increase accordingly.

But data centers also force us to move beyond slogans.

Zuckerberg describes Meta’s facilities as being among the world’s most water-efficient and says the company intends to become water-positive by restoring more water than it consumes in the watersheds in which it operates.

Meta’s sustainability reporting says its restoration projects returned more than 1.6 billion gallons of water to high- and medium-stress regions in 2024. Its projects can include measures that improve watershed function and water availability.

I commend that work. I also think communities should understand exactly what the words water-positive mean.

Restoring water through watershed projects is meaningful. It is not the same statement as saying a data center consumes no water. Efficiency is also a relative measurement.

A facility can be remarkably efficient per unit of computing power while the total scale of computing grows so quickly that overall resource use continues increasing.

Energy illustrates the issue. The U.S. Department of Energy reported that data centers consumed about 4.4% of American electricity in 2023 and projected they could account for roughly 6.7% to 12% by 2028.

That does not mean we should stop building them. It means communities deserve transparent accounting of both the benefits and the costs.

And perhaps we should ask whether more of this infrastructure can be intentionally directed, where technically feasible, toward rural and economically struggling communities that desperately need new tax bases, investment, training and infrastructure.

At the same time, we should respect the residents who say no. Not everyone wants his hometown transformed. Not everyone measures prosperity in tax revenue.

Some people value quiet, land, familiarity, family and continuity more than economic growth.

They are not defective because they are satisfied.

The future may belong to everyone.

That has to include people who do not spend every morning pursuing the next version of the future.

What happens when perfection becomes the expectation?

Near the end of Zuckerberg’s essay, the idea that stayed with me most was recursive self-improvement: AI systems improving their own capabilities, potentially finding ways to wring dramatically more intelligence from the same computing resources.

Technologically, that is extraordinary. Philosophically, it brings me back to where I started.

Where does the pursuit of perfection end?

Companies optimize their AI.

The AI optimizes the company.

The company forces competitors to optimize.

Workers optimize themselves to remain competitive.

Students optimize themselves to enter the workforce.

Our personal agents optimize our health, calendars, spending, relationships, exercise and hobbies.

And then the systems optimize themselves again.

At what point are we improving human life, and at what point are human beings reorganizing their lives around the demands of optimization?

Will a child still spend three hours trying to understand something simply because the struggle itself is part of learning and maturing?

Will curiosity survive instant answers?

Will we still hold a book in our hands and read it from beginning to end when an agent can give us every important idea in thirty seconds?

Will imperfection remain something human beings are permitted to enjoy?

I do not know.

And I do not think Mark Zuckerberg knows either.

That is not criticism.

Maybe admitting that none of us knows is one of the most important safeguards we have.

The future is for everyone. So is the responsibility.

I finished Zuckerberg’s essay more optimistic about some things than when I began.

I want a private personal AI agent.

I want scientists using artificial intelligence to cure diseases we cannot cure today.

I want a small-business owner without Silicon Valley capital to be able to build something world changing.

I want the best real estate agent to become exponentially better.

I want children whose parents cannot afford private tutors to have access to extraordinary education.

I want American AI to lead the world.

I want open systems to prevent a handful of institutions from possessing all of humanity’s intelligence.

And I commend Zuckerberg for being willing to publish a philosophy rather than merely a product roadmap.

But I am not ready to accept that distributing extraordinary power is itself an adequate safeguard against extraordinary power.

I am not convinced competition always creates balance.

I am not convinced job creation somewhere in the economy solves displacement somewhere else.

I am not convinced billions of people trusting an AI agent proves that its aggregate effects are aligned with humanity.

I am not convinced efficiency equals quality of life.

And I do not believe ordinary people should have to become extraordinary optimizers simply to earn the right to remain economically relevant.

Maybe the most important question is not whether superintelligence should be centralized or distributed.

Maybe it is:

What kind of society are we trying to preserve while we build it?

Mark, I hope you are right.

I hope the coming decades bring extraordinary abundance.

I hope people live longer, healthier and freer lives.

I hope small businesses flourish.

I hope inventions emerge from people who could never have afforded to build them before.

I am betting a lot of my own future on many of the same assumptions.

But those of us building with AI—whether we command hundreds of billions of dollars in infrastructure or run a small company in South Carolina trying to make it big—should be willing to ask ourselves the same uncomfortable question:

Just because we can make people more powerful, faster and more capable, does that automatically mean we are making their lives better?

And if the answer is sometimes no, then freedom does not require us to abandon boundaries.

It requires us to build the right ones.

The future should be for everyone.

So should the responsibility for deciding what kind of future it becomes.

About the author: Tiffany Johnson-Wilson is a South Carolina entrepreneur and founder of Johnson & Wilson Companies. She is developing AI-enabled technology intended to change how consumers and professionals navigate real estate. She also serves in local government; the views expressed here are her own.