Agentic AI is Ending Human Agency

and the world we know

The recent phenomenon of ClawdBot and so-called vibe coding has exposed an inversion that many people didn’t see coming; the bots are renting access to humans. Tech start-up RentAHuman has announced “AI needs your body,” and is offering a service similar to the Gig Economy’s stand-by personal services industry. Humans are giving autonomy to their AI Agents, and with that autonomy the Agents are finding the need to have access to the physical world beyond their current means. The answer to some, it seems, is to rent a human to perform the task of manipulating the physical world from the digital one. However this is only temporary as Elon Musk is busy training armies of Optimus robots to be able to perform any and every task that humans do in everyday life.

This is a prime example of the inversion of reality that is taking place before our very eyes as we move all of human life into the digital world. What does that mean for that long treasured prize that has been enshrined in our very founding Constitution that we call human agency? Let’s dive in.

The Current Dilemma

Consider how the paper “Agency: What Does It Mean to Be a Human Being,” describes the concept of human agency: “The most basic conception of human agency is “meaningful, purposive self-direction.” More casually, we may know this as “free will,” or culturally as the inalienable rights of life, liberty, and the pursuit of happiness.

The paper expands on this with the summary, “Ultimately, what the analysis presented here means is that to be human is to be creative, and truly and always open-ended, to be the very site and source of possibility, purpose, and meaning.”

Now this is a more scientific description while our common understanding of Agency is more deeply rooted in the Christian definition, but we will use this for now.

For us, realizing purpose and meaning has always been tied to how we decide to live and how we act out those decisions through what we do; what we will do with our life, why, and unto what purpose. The Bible says that faith (what we believe), without works (what we do), is dead. Aside from the theological implications, this is also simply a truth of being. Belief and action are intertwined and are the foundation of agency. But free will in every context has always had limiters. As people born in a certain place and time and context, we have choices within frameworks. We do not live in a lawless society. We do not have endless means. We each have unique talents, abilities, gifts, and also challenges and disabilities. But we live to serve a higher purpose and to find fulfillment and meaning in life as an expression of who we are. For Christians, we live to worship God and all of our free will is bound to that.

Also presented in the aforementioned paper is analysis of some of the various external factors that limit our opportunities, noting, “Human agency is, however, quite helpless, or at least hapless, in the face of powerful invisible, causally determinative forces. The more such causally efficacious constructs or causes there are, and the more arcane they are – insofar as they are available only to the intelligentsia or supermind – the less agency and freedom there is available to the mass of humanity.”

In other words, the greater a system of controls we live within, the less agency we have, and the less agency we have, the less opportunity to find purpose and meaning, and the less we have of that, the less meaning there is to be found in being a human. Without agency, we lose humanity. For Christians, without agency, we lose worship.

The AI Revolution presents us with an immediate problem, because it plans to move all our current activity to machines via Agentic AI; thus granting the machines increased Agency while threatening to diminish ours. So what is the relationship between man and machine in the AI world to come?

“Humanity is a biological bootloader for digital super intelligence.” “Intelligence that is biological will be less than 1%.” “In percentage, almost all intelligence will be digital.” “Consciousness will be multi-planetary.” “None of us will have jobs.” These are some of the visionary words spoken by Elon Musk as he has proclaimed his expectations for the results of the AI Revolution here, and here. This paints the picture of a diminished and almost irrelevant role of humanity in the AI-powered world. Humans for rent, rather than in charge.

Now consider an excerpt from C.S. Lewis’ work, That Hideous Strength, which reads almost as a direct support for the vision conveyed by Elon Musk:

“That is the point. In us organic life has produced Mind. It has done its work. After that we want no more of it. We do not want the world any longer furred over with organic life, like what you call the blue mould-all sprouting and budding and breeding and decaying. We must get rid of it. By little and little, of course. Slowly we learn how. Learn to make our brains live with less and less body: learn to build our bodies directly with chemicals, no longer have to stuff them full of dead brutes and weeds. Learn how to reproduce ourselves without copulation.” - Professor Filostrato

Perhaps we can easily relate Filostrato’s ideas to our present day as we are engaged in making our digital brains less and less dependent on human bodies, even reproducing ourselves with Agents and robots rather than people, and replacing that blue mould with data centers.

In fact, this is already happening as 2025 and 2026 are witnessing continual waves of widespread layoffs attributed to Agentic AI in the workplace that has replaced the need for humans. News of this happening is usually interpreted as “bad news.” When we hear of jobs lost to AI, we perceive that as a negative outcome. I suspect this natural aversion is why many people who justify their use of AI try to claim that it doesn’t have to lead to job loss, but can also lead to increased potential. Not an end, but a change. Well, we will see about that.

Still, that qualification is important because we seem to naturally recognize that replacing people with machines is inherently bad. And it is good that we recognize that, but that moral judgement is also in direct conflict with the new morality we are expected to adopt in the AI Revolution, because 100% unemployment is in fact the goal.

What will that mean for us as individuals who occupy this world and through our activity find our being and meaning? If all the work is done for us, all the decisions made on our behalf, and the comings and goings of life as we know it delegated to machines, will we have any say anymore in what our pursuit of happiness will entail? If human activity becomes the domain of Agentic AI and of humanoid robots, does that mean we lose our agency?

If everything is in fact orchestrated for us in an ecosystem of total automation managed by machines, what decisions will be left for us to make, and what opportunities will we have to act? Will the nature of the careful orchestration of human life prevent us from exercising free will for fear that any one of us might crash the system for all of us? Will we ourselves present an unacceptable risk to the greater ecosystem by having for ourselves the power and opportunity to exercise the super intelligence we are promised?

In my paper titled “An Ethical Framework for AI,” I asked us to remember that AI is a tool and as with all tools it should only be used according to its designed purpose, by those who understand and are prepared to use it accordingly, for the ends to which it is designed, and ends that are aligned with our framework for morality - for what is good. I also called into that framework the simple example of a moral framework found in the medical community: to do no harm. But before we can discuss the applicable morality of any use of AI we have to remember that human agency - free will, pursuit of happiness, worship - supersedes it all.

If we are indeed transferring agency to machines, can good come from that, or will it simply lead to slavery from lost agency?

In this paper we are going to leverage that moral framework to review the topic of agency and how the adoption of Agentic AI may be undermining our agency as individuals; how it is making our new brains (AI) live with less and less body (humans) and therefore our world needing less and less of us and having less and less room for us; how the goal of Agentic AI is to reproduce ourselves (our activity) without copulation; and that the greater risk from the cumulative effect of Agentic AI adoption is the undermining of humanity itself, and thus, causing harm by its very nature.

Our concern is that moving our human activity to machines may be a form of or perhaps the means by which we will give up human agency altogether. For what we do in the world through ourselves is an expression of who we are, but if we limit what we can do, we will be limiting the opportunity to both realize and express who we are. We will be giving up that by which we find meaning and we will be giving decision making and authority to automated orchestration by machines, called upon perhaps only to touch grass for them. This could represent a stifling of potential so that whatever exists within us cannot manifest or be realized fully outside of us. For if we cannot express ourselves according to our self interests and abilities, then we will have lost our agency, and if we lose agency, we may lose what it means to be human.

Some have argued that with Agentic AI they are simply changing what we can do, not ending human agency itself. Our concern is that a world wholly delegated to the automated orchestration of life by machines is a world of limited options for humans as far as we understand our purpose and nature today. And since this AI Revolution is being built by voluntary adoption among people, our participation is a form of giving ourselves away, and as some give themselves away by building their AI replacement, they also deprive others of opportunity and ultimately agency. To convert your work as a Software Developer to an Agent so that anyone can use the Agent to do software development, is to end the opportunity for your peer Software Developers to pursue meaning and purpose through that activity to which they may be gifted and called.

We propose that Agentic AI is the end of human agency, and so this Revolution is an end to humanity.

A Jobless World

The AI Revolution’s promise is that, “none of us will have jobs.” While Sam Altman and Elon Musk have softened this tone a little by admitting that we may have hobbies in the future, it is clear that they expect robots will perform all the human actions of today.

To a lesser extent, Dario Amodei who leads Anthropic seems to prefer that a few of the most elite will still have jobs (“available only to the intelligentsia”) as orchestrators of armies of bots that do the work that humans once did in order to provide for us and keep society functioning in a practical sense.

Regardless of the degree of job loss, the vision that has been set before us is to replace all or nearly all current human activity with machines so that all our needs are provided for and so we are released to do something else. Sam Altman has expanded on this with the hope that technology will not only provide for us on demand, but will also anticipate our needs and will automatically fulfill them as needed before we even ask. Indeed he has hinted that he’s working on a smartphone replacement device to this very end.

Elon Musk is currently building robots that are being trained and engineered to perform anything and everything that humans do. He has regularly posted videos to X showing his Optimus robots performing household chores, and though people have often quipped that “we’ll still need plumbers,” there is a robot being trained for that human activity as well. Recently, a community of houses built entirely through AI-powered automation was finished as a proof of concept that builders will no longer be needed. The list goes on and on.

When all our jobs are gone, what we will do with ourselves and our time, these Revolutionaries say, is beyond our ability to imagine today. We are to figure that out once we have ended all we do today, but ending all that we do today is the first priority.

Yet if we end all that we do today, we will have stifled “meaningful, purposive self-direction” or at least qualified it in a way yet to be discovered. If we live in a context so governed by determinative forces, can human agency even exist? If it does, it will be far less than what we have today and imposed by our current will of surrender which will imprison us in the future.

The Revolutionary answer to that is of course the promise of transcendence; that by giving up our current activity, we will replace it with potential beyond our current comprehension. As Sam Altman projects, each of us will have the collective intellectual capacity of all of us, and so there will be no limits to what we will be able to do (so he dreams). But first, we must exchange ourselves for AI Agents, and accept that we will learn of the consequences or potential, later on.

This is why it is important to understand, as I have written in another paper, that the goal of 100% unemployment has significant consequences. While that final state still seems very far off because it will of course require robots among us and autonomous vehicles everywhere, the immediate path toward those outcomes is Agentic AI. It is the means to the end. The first step. The fact that the robots are calling for humans to manipulate the world for them means we are well underway.

Agentic AI

Agentic AI is essentially creating a purpose-built micro instance of AI in order to autonomously carry out tasks based on variable inputs from the environment around the agent. An individual agent has the technical architecture of a larger AI model scaled down to micro size, but it is scoped with a limited set of information and instructions that gives it a more narrow purpose. Typically AI Agents are one of many that work together as part of a larger ecosystem; similar to an individual worker trained to do one task among many within a factory. They are usually task or purpose oriented and contained to an identity and purpose. You probably know many of these agents by their names Companion, Worker, Advisor, Support Agent etc.

IBM describes the basic role of an AI Agent as to autonomously decide what to do, and to do it. Agentic AI, they say, “consists of AI agents—machine learning models that mimic human decision-making to solve problems in real time.” IBM also cites an example saying “Agents can, for example, not only tell you the best time to climb Mt. Everest given your work schedule, it can also book you a flight and a hotel.”

Decision making and action automatically orchestrated by the machine on behalf of humans. Not only decision making, but “the best” and “optimal” decisions for us as well as automated actions that carry them out in effecting real life. That is human agency divested to machines.

When building an agent you provide it a set of information that provides it context from which to operate, and then you give it a set of guidelines by which it will operate in response to new input. Many people even give their agents unique names.

It is common to convey this identity to an agent by writing what is called a training prompt for the agent, as if you are telling a new employee who they are, what they need to focus on, what you expect them to do, and how you expect them to do it. In fact, in the workplace, agentic AI is often referred to as AI Workers, and they are being used to replace people. So in effect, training an AI Agent is training a digital employee. In a recent Interesting Times interview by Ross Douthat, Claud AI CEO Dario Amodei described Claud’s training manual as like a constitution that tells it what it is, why it exists, what it has access to, what it should and shouldn’t do etc. It is eerily similar to making a new being; you are speaking meaning into an “agent” and giving it decision power so it will operate autonomously on behalf of humans.

The agent is thus prepared to decide what tasks to execute and to perform them when certain input is received. It can deliver the results back to itself to perform a new iteration of work, or it can pass the output on as a completed task. It is designed to be self-deterministic to a degree, but to choose what is optimal, and to execute that autonomously without human attention.

In common use outside of the workplace, Agentic AI is often used by people to perform tasks or to give them the ability to perform tasks instead of hiring a professional to do the work they cannot or do not want to do. Agentic AI is also used commonly to simulate experiences and even people, notably through the creation of AI Companions, AI Teachers, AI Assistants, AI Advisors etc. Each of these has a context that describes the role the agent plays for whomever they belong to, but all of them are examples of replacing people and activity with machines. That’s replacement and a limiter on human opportunity.

While you may see the main purpose of an AI agent as to automatically perform work on behalf of humans or to do work that humans cannot (as is the classic use case of robotics), practically speaking Agentic AI is mostly being used to replace humans by transferring their work. This is the celebrated goal of the AI Revolution as that transfer facilitates the dream that “none of us will have jobs.” In fact a key metric of success used by AI Revolutionaries is that human participation in work is being eliminated: i.e. through layoffs.

The first version of Agentic AI I worked with professionally was designed to understand the structure of the code my team of Engineers maintained, and to automatically author additions to the code to address new use cases that we would recite to the agent via its prompt interface. In classic AI workflow style, we entered our desired outcomes and it created a representation as code. The goal was to transfer the work of writing code to an AI Agent. Once that was built, the creative part of our role was replaced by lists of use cases obtained from outside sources. The final result of this was that our team was eliminated - allegedly replaced by AI Agents that could read lists of use cases created by others or by themselves, and translate them into code updates that were inserted into the software product we maintained. This was celebrated by executive staff in a press release sent to investors that announced hundreds of jobs eliminated thanks to AI.

What did that mean on a practical level to the humans involved? I and my coworkers were no longer free to engage in the work we chose and from which we realized meaning. We found the balance of our skills, interests, and needs fulfilled through the technical research and engineering work we did. We enjoyed coming up with use cases and solving the technical challenges of integrating them into our code, and we enjoyed the results of seeing our work produce positive outcomes for our customers. But all this opportunity was denied in the name of efficiency. But as the overall industry is doing this same outsourcing, I and my coworkers no longer have the opportunity for that self-determined application anywhere. Our pursuit of happiness has been removed as an option from the table of choices in front of us. But in my case, this opportunity was the result of decades of a professional career. It’s not just that I wanted to pursue this individual job, but that this is where my life and calling had led me.

Denied in favor of efficiency.

In recent days, the ClawdBot phenomenon has expanded Agentic AI into the software development world as people “vibe code” agents that perform everyday tasks for them from controlling the IOT devices of their house, to managing their finances, or replacing software they bought elsewhere, or any number of other life-tasks that can be automated with technology. As those use cases are expanding into touching the real world, those Agents are now asking for bodies. Elon is ready to meet that need with his robots.

Some other example of vibe coding included an agent that can analyze your tax returns to find new exemptions and lower your tax bill (ending tax advisors), an agent that can analyze the attributes of a property and determine its potential value so home purchasers can submit a bit (replacing real estate agents), an agent that replaces DocuSign with a free version of similar features, and even an agent that will help you write the optimal prompt for your other AI agents.

But to recap, Agentic AI is about using a purpose-built AI Agent to perform the work or tasks of humans, so that people no longer do that work. It is the transfer of human activity to machines.

Some say that this will empower humans to do more with greater efficiency and to finally get to more important things that can’t be done today because we are busy with the mundane. Indeed, that is one of the main selling points Sam Altman likes to cite. While I have argued previously that we are already doing the most important work, and while no one can yet say what the more important work is, Altman simply defers that question to the future; we’ll figure it out later. But more importantly, by ending the work of people today we are defining their future by denying them opportunity. This is undermining human agency.

Human Agency

Human agency is of course a philosophical or metaphysical topic that is certainly formed by theology and culture. But in simple terms it is the essence of being you and choosing for yourself what you will do with your autonomy. Having agency means you decide how to organize, manage, and conduct your life, and that you can act according to your will unto finding meaning and self-realization in life. You get to decide what you will do that is meaningful to you. You have control over yourself. Note the similarity in IBM’s definition of agency in the form of AI.

Culturally we may recognize our free agency as “the pursuit of happiness,” which is made possible by your having been created by God and thus bearing intrinsic value in the world and the right to pursue the life God has given us according to His calling and purpose. Biblically we often recognize this as free will. But perhaps even more simply put, having agency means you decide what will be done in your life, and then you act on that decision, thus realizing purpose and meaning and fulfillment through that.

As a more practical example, a “free agent” in professional sports is a person who is currently unaffiliated with a team and is open to finding a new one to become part of. They get to pursue different teams and negotiate the terms of joining one. Once they sign a contract with that team, they are bound to the terms and are no longer a “free agent.”

Conversely, lacking agency means you have no choice, you simply execute the orders of others or based on natural impulse, or you simply live by the very limited number of choices before you, likely none of which are preferred.

However, part of having agency is the responsibility to use it well and in ways that are morally good. Free will can be exercised to commit harm to yourself or others, but it can also be used in the expression of abstaining from committing harm even when the capacity to do so exists. This is why it’s important to maintain full agency - deciding and acting - but also within a framework that enables and protects everyone.

As human beings of Western heritage, we hold our individual agency as a great prize to be cherished, protected, and used responsibly. Freedom and liberty are in our political ethos.

In the Christian church, we seek to balance our free will with the sovereignty of God which is also reflected in relationships we have with others in our family and in our larger Christian community. Indeed as willing participants of a shared society, we also balance our individual agency with that of others so we can exist peacefully in cooperation with each other toward our collective good will. We do willingly give up our agency from time to time in service to each other and in obedience to God. But this is part of our self-determination toward meaning. These are terms we accept through peaceful participation within our society and culture. It is part of the social contract we have with each other. But it is by our free will.

In other parts of the world and in non-Christian or anti-Christian worldviews like marxism, individual agency doesn’t exist. Society holds a collective agency and every individual’s responsibility is to contribute to the needs of the many by sacrificing their personal self-interests, and even their personal needs. Your personal interests don’t matter. What matters is that you serve the individual role assigned to you by the orchestrators of the greater society. Individuals are grown, not unlike training an AI agent, to be equipped and prepared to serve a role when called upon as part of a larger ecosystem of peer agents.

This is why we can see the very obvious parallels between the AI Revolution, slavery, and marxism or socialism. We are building an army of defined agents (workers) who will do everything at our bidding, but in such a tightly orchestrated and interdependent ecosystem (all of life) that mistakes cannot be tolerated and so free will cannot exist.

All the agents must work together and they cannot work outside of their designed purpose and code. What if we humans want something else? Opportunity in socialism is denied.

Divesting Agency

Returning back to Agentic AI, one of the dangers individuals face with using this technology, is divesting their personal agency to their digital agent which can then become a master they are bound to. It is incremental and small, happening app by app or bot by bot. As this grows across a society, we become less and less free and increasingly bound to the machines that now orchestrate life on our behalf. If we are no longer doing the activity of life, then we are limited by those who are.

While this does not mean that using AI means you are always divesting your agency and therefore losing yourself, using agentic AI can lead to that end and indeed we are very aggressively encouraged to use the technology for that purpose, for that is what it is designed for. That is the current which is driving us. The pretext is that these Agents will be our masters and our slaves.

There is no clearer indication of that than when Elon Musk said “probably none of us will have jobs” and that his Tesla robots will provide everything we need - doing everything for us. The outsourcing of all human activity, or the divesting of all human activity to the machines is the final goal he and the other AI Revolutionaries are working toward. Agentic AI is the incremental progress toward that goal. As that system of orchestration of life grows into its total form, we will of course be limited in our decisions and opportunities and thus our agency.

This is most obvious in the workplace when a person is laid off and replaced by AI, especially if their skills make them no longer employable because all other companies who could employ them are switching to AI instead. They are effectively no longer able to exercise their agency in the world, or at least it is being severely limited if it is by that work that they found purpose and meaning in life. But that’s not exactly giving up your agency directly or entirely. Indeed, that may be worse because it is a form of you being denied your agency by a society that is changing around you as it gives up its agency little by little and yours is caught up in that.

An example of giving up your personal agency can be found in Gemini product marketing in which users of that AI agent are invited to use it as a Travel Agent to plan their vacation end-to-end. This is a use case repeated in a viral video posted recently on X, which showed a man doing something similar via his Tesla Cybertruck and Grok. With a vacation planning agent, you can simply speak to the machine and tell it your basic criteria, and the machine will make the plans and the arrangements on your behalf, simply telling you what you will do, when, and where.

Now, of course today you retain final agency in that you decide whether or not you are going to do what the AI agent planned for you, but you divested or outsourced the planning process to the machine. That’s the first step. You gave up the work of pursuing your vacation itinerary for the sake of efficiency, and the point of the agent is so that you no longer have to do that. That’s part of the Agentic AI marketing; to release what you would do to the machine so you can do other things with your time. If you simply followed through with the AI generated vacation plan, then you would have given the machine agency over you. You gave the machine, the agent, the authority to decide on your behalf and then you simply followed those decisions and experienced what the machine planned for you to experience. When combined with a self-driving vehicle and credit card information stored within the agent, the AI Travel Agent can even act on its decisions and take you along for the ride, literally.

In addition to divesting your own agency in this situation, you also facilitated the termination of people who work as Travel Agents, compromising their agency. That may not be true if you were the only person who did that, but if everyone switched to automated travel agents, then human travel agents would cease to exist.

Again, Agentic AI is designed for this purpose - so you can delegate to the machine meaningful and important tasks that are your responsibility that it will do on your behalf. Usually this is pursued in the name of efficiency - to save time or reduce complications. Rarely it is spoken of as an opportunity for you to explore greater potential and opportunity - because it is a limiter by default.

But what if in the AI world no one can drive anymore because all vehicles are converted to self-driving cars so full orchestration of human activity can be maintained efficiently and safely? Then again, one of those most cherished of all expressions of agency, driving, will be delegated to machines.

We might be tempted to think this is the same thing we have been doing with technology and automation for years, however, giving up agency is not the same thing as using a tool to aid in your work or activities, which is what automation and technology has typically been used for. Agency is more about who you are and what you decide to do with your life and losing agency is where you lose the ability to exercise free will or are bound to the outcomes decided outside of you. But that outcome is the cumulative effect of all of this outsourcing.

The AI Writer agent is a great example of self-replacement. But more importantly, by developing AI Writers, we are limiting the opportunity for agency itself for all. A common use case for this AI agent is for people to author a written work in draft form and then upload it to the agent asking it to improve on the work for them; re-order, re-word, simplify, expand, find sources etc. Whatever the AI agent produces can be incorporated into the final work, which means it is no longer wholly the expression of the author, nor a reflection of their skills, nor of their decisions. It is now a co-authored product. Less of the author.

Part of learning to be a good writer is to practice and develop writing skills yourself. To work and rework your writing until it is the best you can produce. To become a great writer in pursuit of purpose and meaning and the full realization of your potential according to your gifts and calling. The process itself is part of exploring opportunity and realizing meaning. The work has tremendous value and indeed may be the very point of it all.

By delegating that to AI, you are letting that go and transferring authorship to the machine, asking it to serve in the role of a writer and an editor. That also compromises the agency of another person because the role of editor might be naturally served by a person gifted and well practiced in that skill. But with accepting edits performed by the agent you are letting it decide what to write for you; compromising both decision making and action and also making your work no longer your own, and stifling opportunity to learn and grow in the process. Your final product is now bound to what the AI agent produced, and you deprived yourself of the opportunity to expand and realize your full potential. You will have settled for less meaning.

In the future, a reader of your work may ask you “why did you phrase it that way” or “what made you organize your points in that sequence,” or “can you mentor me?” The truth is, you weren’t sure what to do, so you delegated that to a machine and let it decide. Now you are bound to that, and that means you are bound to maintain or support that work by another, rather than a work that could have been naturally your own. Depriving yourself of the process also deprives you of the benefits and opportunities that may have elevating meaning.

But again, as we collectively build AI Writers and use them and circulate the works produced by them, then we are replacing writers among us. We are ending opportunities, and that limits agency for all.

Another example of releasing personal agency may come in the form of an AI Agent I reviewed that claims to be a “Career Advisor.” Now in the traditional world, a school guidance counselor might serve this role for you, but choosing your career and where you will work in that career is a highly prized part of expressing our agency. With the Advisor Agent, you input your preferences, goals, dreams etc. and it will generate a bunch of content to guide you toward realizing those goals. This can include telling you which educational courses you need to take, from which universities, as well as what companies to apply to and for what roles, and it will even fill out applications or write resumes for you. It even advertises as being able to tell you where you should live in order to access the most opportunities in this pursuit, including salary ranges for jobs in those locations. This is far more than advising. This is facilitating.

If a young adult were to use this Agent to the full potential of what it is designed for, then they would be wholly divesting that part of their agency to the machine. They would be letting the machine decide for them where to live, where to go to school, what classes to take, what company to apply for etc. ending all the natural progression of that activity by which we change and learn and grow. Again, they could choose to use this simply for informational purposes, but clearly that is not what it’s designed for. It’s designed to do the planning and deciding for you, leaving you to simply act on its orders. It will even do some of that acting for you as is the purpose of Agentic AI according to IBM.

Another example comes from the product vision of Sam Altman for a new form of AI Assistant. He has described a device that will be with you at all times, constantly listening to everything happening all around you. Based on what it “hears,” it will attempt to preempt your needs and facilitate whatever it is it thinks you want or need to do or know. This is clearly releasing management of your life in an intimate and personal way to a machine.

AI Companions are an example of both giving up and taking away agency. With an AI Companion, many people give up their agency because they defer their reasoning and decision making to the analysis and interpretation of their Companion, sometimes unto committing personal harm and crime. They do this because they believe the narrative that AI is a super intelligence and can and will know more than they will and can decide better outcomes for them in life planning and decision making. That’s what Agentic AI is, right?

AI Companions can also take away agency from others when they are used to replace friendships or relationships. We have seen several examples of this among parents who say their AI Companions give them guidance on how to best manage their family in replacement of their spouse. There are even digital health products that offer AI-powered guidance for parents regarding how to talk to their kids about online safely.

There are even AI Companions that replace Pastors, speaking so-called prayers in the voice of the Pastor based on the prompts given by the user. Recently there was word that a church has integrated an AI Jesus into their confessional liturgy. And of course there are Pastors using AI to write and test sermons. All of this is divesting; ending.

In a recent post to X, a person remarked that they enjoy using AI to “organize their thoughts” and another mentioned that their AI Companion can interpret and notice things about them that no one else seems to notice - insights that are helping them better understand themself and that helps them act differently in relation to others. These are more difficult examples of divesting part of personal agency to AI agents because these individuals are using AI as a sort of part of their mind - helping them with decision making. But whenever you are invoking an AI Companion to “tell you” and decide for you or guide you, then you are divesting decision making to the robot and the point is so that you will act on their decisions rather than your own, because you find your own agency lacking.

Rather than journaling or writing down thoughts and ideas and observations and reviewing them on their own, these people are asking AI to do that for them, and then acting out upon the analysis provided to them. It is a form of outsourcing introspection. In fact, a promoted AI agent created by a so-called vibe coder is designed for a similar purpose. You speak ideas or hypotheses to the agent and it analyzes them and then formulates a plan to improve, test, and even implement the idea.

There are many other problems that are related to these practices, but I think at the most simple level they are subtle examples of where our agency begins to slip away through our willful delegation of meaningful activity that should belong to us.

Agency Lost or Exchanged?

These examples of incremental transitions of activity to machines is a gradual process of losing control and subjecting our self-determination (agency) to the options left available to us. Though this is gradual, it can take leaps forward depending on how much progress is desired by the individual. But the real danger is the growth and expansion of that incremental progress until it replaces critical aspects of our life or the majority of our lives so that we are no longer in control. If we no longer have options to decide or the ability to act on those decisions freely, then we will have lost agency.

Now, personal agency again is about choosing for yourself and acting on that choice; reasoning toward the action you find meaningful. In all the examples we reviewed in this article, there is always a way out, at least still today. The person using the AI Agent could have chosen to use it only as an advisor and they could retain the final decision making power and they could choose to act in a way that incorporates the output in part but not in whole. That might skirt the issue of losing agency, but again, the point of Agentic AI is to end human activity. That starts with personal activity and can expand to the point of total dependency on the collective will (or code) of the technology.

In my professional experience, when we offered Agentic AI as an aid to the work of humans while still retaining control over research, use case development, and integration testing (including quality controls), we were told that all of it needed to move to the machines. That’s the goal of the Revolution. Incremental progress is setting the precedent for the next increment.

As we divest our activity to the machines, we lose choice. We lose reason. We lose meaning. If we lose those things, then what is left of our agency?

The AI Revolutionaries driving this transformation offer that in exchange for giving up everything we do today, including decision making, we will gain something vastly greater and vastly different. They offer that the AGI powered world will be so different that we cannot today perceive what it will be; what we will do, what human existence will mean. Sam Altman is very open about this. He doesn’t try to speculate but simply offers glimpses like his claim that in the near future, each of us will have the collective intellectual power of all of us combined. Who knows what we will do with that. His theory is that it will look something like The Burning Man festival where Altman says he saw his vision realized. He proposes that in the future we will use our vast powers individually for indulgence and good will. We will all just do good things with the tremendous power we have. There will only be good, as he defines it. A world of humans with god-like power who will only do good in the pursuit of pleasure for themselves and in cooperation with each other. In other words, he sees an everlasting orgy, or something like that. An everlasting world of perpetual feasting. That sounds familiar.

Elon Musk seems to agree, often speaking of abundant wealth, meaningless money, provision of whatever you want by machines that are capable of meeting our every need and desire.

Alas, during the period of the Industrial Revolution, concerned citizens worried that moving work and life out of the home and into cities and factories would wreck the continuity of life and family. The factory owners promised wealth and wellness. They promised a new era of prosperity and comfort. In many ways they did deliver on that promise. We have seen the general human condition rise to a material and physical richness unlike any in the history of mankind.

But what of our souls? What of the deeper things that matter most? What of our preparation for eternity? What of the state of our minds?

And the Industrial Revolution did not seek to end everything. It was limited in scope and in outcomes. It did not consume everything. The AI Revolution asks to consume all the work and activity of man. It’s an all or nothing game. We give it all in exchange for the promise. It cannot work in part.

We have the benefit of retrospective analysis of the effects on humanity from previous revolutions, and we should learn from them, and count the cost of this one before we proceed. We may lose far more than we gain, and we may not gain at all.

For what if in the process of transferring ourselves to the machines, we lose everything that was worth living for?

That’s the risk. It’s an end without a better beginning. The risk is the end of humanity as we know it. Ironically, that’s the promise as well.

I offer that if the AI Agent undermines your agency or that of others, then we should not pursue Agentic AI as a replacement for all. We should limit it within the framework of our history, tradition, values, and in pursuit of our individual happiness. We should hold human agency as the prize to be dearly protected, at the cost of the promises of the AI Revolution.