Human + AI Lab: Seeking Effective Models of Human–AI Cooperation

Artificial intelligence is developing remarkably quickly.

New models, tools, and capabilities appear every month.

We can generate text.

Analyze documents.

Create images.

Write software.

Organize data.

Hold conversations.

Automate processes.

But at the Vis in Unitate Foundation, we are interested in a question that, in our view, is even more important than the capabilities of the technology itself:

What should effective cooperation between people and artificial intelligence look like?

This is the question around which we are developing the idea of:

Human + AI Lab

The Laboratory for Human–Artificial Intelligence Synergy.

Not a technology laboratory. A cooperation laboratory.

Human + AI Lab is not intended to be yet another place that showcases the latest AI tools.

Technology changes too quickly to build an initiative exclusively around specific applications.

We are more interested in something lasting.

Methods of cooperation.

We want to observe and test how people and artificial intelligence can divide tasks between them.

When should AI propose a solution?

When should it merely provide information?

When should a person take full control?

How should the result be verified?

How can errors be reduced?

How can we avoid a situation in which a person begins to accept the system’s answers uncritically?

These are the questions that interest us most.

Synergy means more than automation

Automation most often involves handing over to technology a task previously performed by a person.

Synergy is something different.

We can speak of synergy when the combined result of people and technology is better than the result either could achieve separately.

It can be expressed very simply:

Human + AI > Human

but also:

Human + AI > AI

Because artificial intelligence without human context, experience, and oversight also has serious limitations.

What do people contribute?

People possess:

• experience,
• contextual knowledge,
• intuition,
• empathy,
• values,
• responsibility,
• an understanding of social relationships,
• the ability to define goals.

It is the person who should answer the question:

“Why are we doing this?”

What does AI contribute?

Artificial intelligence can contribute:

• enormous analytical speed,
• the generation of many alternatives,
• the processing of large amounts of information,
• the automation of repetitive activities,
• help in identifying relationships,
• support for the creative process.

AI can answer one question very quickly:

“How can we do this?”

But it does not always understand whether we should do it.

That is why we need new working methods

Human–AI cooperation should not be reduced to entering one question into a chat window.

It is a process.

A good model might look like this:

1. A person defines the problem.
2. AI helps analyze the situation.
3. AI generates possible solutions.
4. The person evaluates them based on experience and context.
5. AI helps refine the selected solution.
6. The person verifies the result and makes the decision.

In this model, AI does not replace the person.

It extends their capabilities.

What do we want to study at Human + AI Lab?

We are interested in practical human–AI cooperation in different environments.

Education

Can AI be an individual learning assistant?

How can we use it so that a person genuinely acquires knowledge instead of merely receiving ready-made answers?

Public administration

How can AI help employees analyze information, prepare documents, and communicate with residents?

Where should the boundary of automation lie?

Civil society organizations

Can AI enable a small organization to perform tasks that previously required a much larger team?

Entrepreneurship

How can AI support small businesses and people who are starting a business?

Professional activation

Can artificial intelligence help a person changing careers acquire new skills more quickly?

Creativity

Does AI limit creativity, or can it increase it?

Or does everything depend on how it is used?

We are also interested in failures

Human + AI Lab is not intended to be a place for uncritical promotion of technology.

It is equally important to study situations in which cooperation with AI does not work.

We want to analyze:

• model hallucinations,
• incorrect answers,
• automated bias,
• excessive trust in technology,
• data protection problems,
• loss of user independence,
• situations in which automation reduces the quality of a process.

Because genuine research into technology should show both its capabilities and its limitations.

Human + AI as a skill of the future

In a few years, the ability to cooperate with AI may be as natural as using an Internet search engine is today.

But the mere presence of technology does not mean that people know how to use it.

That is why we need new methods of education.

Not only:

“How do we use AI?”

but:

“How do we think together with AI without ceasing to think independently?”

This is a fundamental difference.

And it best describes the direction we want to develop.

Our hypothesis

At Human + AI Lab, we begin with a simple hypothesis:

the greatest potential of artificial intelligence lies not in the technology itself, but in the relationship between technology and people.

We want to observe this relationship.

Test it.

Measure it.

Improve it.

Then translate the experience we gain into:

• training,
• education programs,
• publications,
• good practices,
• methodologies,
• pilot projects,
• solutions for organizations and institutions.

We do not yet know where the limits of human–AI cooperation lie.

And that is precisely why they are worth exploring.

Human + AI Lab

We are not studying a future without people.

We are studying the future of people in a world of artificial intelligence.