Start from the person

A personal statement from Alexandru Goroh, founder of Life GPS.

The side I am on

I am currently in conversations with very different people — researchers, representatives of companies, public institutions and standards groups shaping the relationship between people and artificial intelligence.

They have different responsibilities and interests. But in many of these conversations, I see the same familiar starting point.

We start with the system, then try to find a place for the person inside it.

With the company and its processes. With the state and its responsibilities. With the product and its metrics. With the model and its capabilities. With the analysis we can already perform and the data that already exists.

This is how it became possible to work at the scale of millions of people. Organisations built infrastructure, accumulated knowledge and learned to coordinate extraordinarily complex processes.

Today, I propose changing the point of departure itself.

Start from the person.

We have learned to see a person as a customer, employee, patient, citizen or user. Now we need to bring those roles back together into one life.

What do they want? What did they actually choose for themselves? What would they refuse to exchange for money, convenience or another increment of efficiency? How much time and energy do they have, and what commitments? What happens to their family, health and future after a decision that looks right inside one system?

Who holds that question across all the systems?

What is all this progress for?

We have built an extraordinarily capable world. Possibilities keep expanding, knowledge becomes more accessible, and tools take on increasingly complex work.

So I want to ask a simple question: what are we developing all of this for?

Where is the human result of that progress?

With so much knowledge and technology available, why does a person still have to assemble their own life by hand from dozens of disconnected parts? Why does the need to provide for tomorrow so often leave no room even to understand what they want tomorrow to look like?

We know how to optimise production, delivery, computing and financial risk. Yet the person is still trying to work out whom to trust, what to choose, where to find the time and how one decision might damage another part of their life.

Why does so much progress still leave the person as the manual integration layer of their own life?

Many existing systems are excellent at what they were built to do. That makes it especially important to see the problem that arises between them.

The bank sees the financial side. The employer sees work. The doctor sees health. The shop sees a purchase. The agent sees an assigned task.

The person has one life across all of them.

Every system can be right within its own task — and the person can lose in the sum.

It is like performing heart surgery while some of the monitors are in another department. Each specialist may be doing the right thing. But the readings need to be seen together, because the consequences meet in one person.

That is the gap I am trying to close.

Who protects the person?

Recent events make the question more urgent.

On 29 September 2026, leaders of major AI companies signed a voluntary safety accord. AP reported provisions for internal controls, independent external reviews and board-level oversight. Much of the checking is organised around the developers' own governance mechanisms. AP reporting

For the developer, an error may become a bug fixed in the next version. For the person, the action has already happened.

Fixing the system tomorrow does not undo an action taken on my behalf today.

I experienced this myself: an autonomous agent began sending new emails on behalf of the company without my approval of those messages. → Read the case

Who, then, protects the person?

Who is on their side at the very moment increasingly powerful systems interact with them, persuade them, offer decisions and begin acting on their behalf?

The state can set binding rules for participants whose power is far beyond that of an individual. At the same time, it is responsible for the country's economy, security and technological development.

What happens when the need for stronger protection meets the fear of losing a technological advantage? What works on the person's side while governments negotiate, companies release new systems and the rules change?

I understand why this race continues. I want to participate in technological progress from the side of protecting the person.

If the state cannot be the person's only protection in this race, what needs to exist alongside them?

Behind an organisation's interface stands an entire infrastructure: data, models, agents, specialists, legal teams, analytics, automation and the accumulated memory of an enormous number of interactions.

On the other side is one human mind. Its attention, time and energy.

Laws, rights, professional help and digital literacy are necessary. But a checklist does not review conditions around the clock. A lawyer is not present at every click. A person cannot continuously hold every permission, previous decision and consequence for the rest of their life in their head.

While the person considers a decision, the system can already be calculating. While they try to recall a previous choice, the organisation can retrieve its history. While they examine one consequence, the other side can process thousands of signals.

Human attention alone is not an adequate counterweight to automated infrastructure.

I do not want to make a person smarter than every system they interact with. I want to give their own mind technological support comparable to what organisations have already built around theirs.

Memory. Context. Models and automation. The ability to keep checking what is happening and bring back the questions that require the person's own decision.

Protection of the person has to become technology too.

It needs to belong to the same technological generation as the systems around them, with a different starting purpose: to work on that person's side.

Even a conversation with AI can now sit alongside commercial offers. Ads may appear for users on ChatGPT's Free and Go plans; OpenAI says they are separate from responses and that conversations are not shared with advertisers. OpenAI's explanation

This makes the distinction between sides especially important: the platform has its business model, and the person has the intention they came with.

Who helps them preserve that intention throughout the chain of interactions?

The architecture I am building

I have brought an architecture built around that question into these conversations.

I came with a working personal core through which we can begin testing an answer. Life GPS.

Life GPS is a human operating system covering life as a whole. Its starting point is the person: their situation, wishes, constraints, relationships and decisions. Security and human control are the foundation of this architecture.

Models provide capabilities. Agents execute tasks. Services, professionals and organisations participate in the parts of life where the person needs them.

The person's own side is designed to hold the connections between those actions: remember the chosen goal, surface conflicts, account for changing circumstances and bring consequences back into the wider picture of their life.

I call one of the protective functions of this side Human Antivirus.

The value of antivirus is that protection becomes working technology. It stays alongside what is happening and performs checks automatically.

We need the same transition for a world of AI and agents.

Is this actually what the person asked for? Is the executor acting within permission? Are the conditions still the same? Does the person need to be asked again? What does this action change elsewhere in their life?

A person must be able to choose, take risks, reconsider and change direction. The protective layer must support that autonomy, respecting applicable law and the rights of other people.

A system that protects a person's choice must itself remain under their control.

There is a fundamental boundary here:

To understand a whole life, we do not need to give the whole life to every organisation.

The architecture keeps the wider context on the person's side. Each external interaction is scoped to the information and authority needed for its particular purpose, taking account of the participants' lawful responsibilities.

We want to connect the world's existing capabilities around the person. Banks, medical systems, public services, shops, models and agents already exist and continue to develop. Each keeps its role.

Meeting in the middle

On the same day, 29 September, the United States introduced America.gov — a unified entry point to federal information and services. At launch it answers questions; transactions are intended to expand as agencies connect their services. Official launch fact sheet

The state is bringing its side of the interaction together. We are approaching from the person's side.

I want our protocols to meet in the middle.

That is why these conversations with organisations and institutions matter so much to me. I do not know every public procedure, every jurisdiction and every industry from the inside. That knowledge belongs to the people responsible for those parts of the world.

I bring the architecture of the person's side, and I am ready to learn from those who know the other side.

Together, we can establish how to connect a person's intention with real possibilities, laws, obligations and constraints. How to verify authority, disclose the necessary minimum, receive a result and preserve the ability to reconsider a decision.

Billions of people have to live together, work, exchange, build families and companies, honour commitments and reach agreements. The person needs a side of their own within that shared system of interaction.

From more possibilities to a better life

For me, this brings us back to the purpose of progress.

I love this world for how much there is to learn, build and try.

I want the quantity of possibilities we have created to become quality in people's lives.

For a person to understand what they want. To see the paths available. To understand consequences. To receive help. To change direction. To find time for people they love, curiosity and life itself.

People will still have different wishes, constraints and conflicts. But each should have more ability to understand what is happening and participate in the decisions that shape their life.

The personal core of Life GPS is already running and being tested by early users on iOS. The external interaction protocol is published for review; external integrations are not live yet. We still have to test the architecture in real interactions and demonstrate that it does its job. There is no integration with America.gov or endorsement by the organisations discussed here.

That is why I invite researchers, developers, companies, public institutions and people themselves to examine it, find its weaknesses and improve it together.

If the world equips every system around a person with intelligence, the person cannot remain without a system of their own.

We are building that side.

If we say that progress exists for people, let us build from the person — and return to the result in their life.

To be on the person's side, we have to start from the person.

That is where I started Life GPS.

— Alexandru Goroh

Source note — 2 October 2026. The AI accord is described from reporting, not an independently reviewed original agreement. The linked OpenAI documentation and US government fact sheet describe their own services. The interpretation and proposal in this statement are the founder's position; these sources are not evidence of Life GPS's effectiveness.

Every claim on this page carries a status