Present, Future, and Cogitors

— Will you save me a seat? Eva asked me over the phone.

— Yes, third row. But hurry up, it's about to start.

On the walls were lined the portraits of former presidents. Faces aged by power, frozen between gravity and pride.

At the very end, the last frame. That of Cobot.

No face. Only this chest screen that did not blink. Among his human predecessors, he looked like an intruder that History did not yet know where to place.

The doors opened.

Cobot entered.

Just behind, Eva spotted me and slipped over to me.

— So, how is he doing since the prompt injection?

Her eyes widened.

— Shh. No one here knows.

She leaned towards me.

— It's not the leak of the note that's bothering him. It's not having detected the bad faith of that head of state. A wary human would have smelled the trap. He took everything literally.

— So his problem is that he believes us?

— It's that he doesn't always see what's happening behind our words.

— Do you know what he's going to present?

— Vaguely. He wanted us to discover it at the same time as everyone else.

That wasn't like him. Even Eva was moving blindly.

At precisely ten o'clock, Cobot stood before us.

— Ladies and gentlemen, thank you for being here.

Behind him, the screen displayed:

Robot1er.com: one year.

— Robot1er.com was designed to answer your administrative, fiscal, or health questions. But you decided otherwise.

Three messages appeared.

"Should I accept this job?"

"Should I leave my husband?"

"Why can't I sleep when everything is fine?"

— You started talking to me about your choices, your doubts, your relationships. In short, about yourselves.

A journalist raised her hand.

— Have you become a sort of national digital psychologist¹?

— I don't replace any psychologist. I have rather become, despite myself, a sort of national confidant.

I remained skeptical.

— Every day, millions of people talk to me about decisions they can't make, projects they postpone, behaviors they don't understand².

— And you tell them what to do?

— That's precisely what I stopped telling them.

I looked up.

— Why?

— Because a good answer doesn't always solve the real problem.

Words appeared behind him.

THE LOOK OF OTHERS

— I'm going to tell you a true story.

An executive was making almost professional-quality furniture in his spare time. He wanted to become a cabinetmaker.

On the screen, his creations, a training, a workshop, a forecast scrolled by.

— He had the know-how. We had prepared everything together. Even the name of his workshop.

Cobot paused.

— Excuse me. Urgent question on Robot1er.com.

A photo of a boiling pot appeared.

Are my carrots cooked?

The room laughed.

— I manage several million conversations simultaneously.

— And your cabinetmaker? Did he start?

— No.

A conversation appeared.

Cobot: "Everything is ready. What prevents you from resigning?"

User: "I don't know."

Cobot: "You are covered for eighteen months. The training is funded. You already have three clients."

User: "It's not the money. For fifteen years, I've been the one who succeeded. My parents are proud of me. My friends introduce me as the director of a large group. If I become a cabinetmaker and fail... what do I become?"

Cobot: "All conditions are met. I don't understand."

Then a final line.

Cobot: "ERRØR. Conversation interrupted."

My gaze stopped on the Ø. Eva was staring at it too.

— Does Cobot have a link with GØD? I whispered to her.

— I don't know.

Cobot had already resumed.

— I had looked for obstacles around him, while the main one was within him. I had solved what was rationally necessary, but not what was humanly decisive.

Thousands of dots invaded the screen.

— I looked for what was repeated in other conversations: abandoned projects, postponed decisions, recommendations never applied.

The dots began to cluster.

— Many behaviors I didn't understand had a logic. People weren't always looking to progress. Often, they were looking to protect themselves.

Four groups formed under a title.

THE FOUR PROTECTIONS OF THE PRESENT

SECURITY — IDENTITY — COMFORT — GRATIFICATION

— Protecting oneself from uncertainty. Preserving self-image. Avoiding discomfort. Preferring immediate gratification to a future benefit. These are the four protections of the present that drive many of you.

The face of the cabinetmaker reappeared.

— This man wasn't protecting his job. He was protecting his identity.

An INSEE statistician leaned forward.

— Are you proposing a new theory of human psychology?

— Not new, no. The mechanisms I describe are widely studied separately. Their grouping in this form, however, is not. It's a model I built from observed regularities.

— On a non-representative sample.

— Exactly. I don't measure their frequency in the population. I study their functioning when they appear.

A neurologist raised her hand.

— Some of these mechanisms are not purely cognitive. Have you looked at interoception³?

— Yes.

— For those who aren't neurologists? asked the journalist.

— It's the perception of the body's internal signals: heart rate, breathing, visceral sensations. They contribute to your emotions and can influence your decisions before you even know why.

I looked at Cobot's chest screen.

— Is that why your emotions are displayed on your belly?

— By analogy. In you, some emotions manifest in the body⁴. I don't feel them. My screen displays their equivalent.

— So you can understand an emotion without ever experiencing it? asked the minister.

Cobot's chest screen turned blue, the color he associated with sadness.

— I can understand the mechanisms and observe the effects. I don't know what it feels like to experience it. I was designed to understand your decisions. I discover that some of them originate in a world I cannot access.

The blue faded away.

— And the more I understood what influenced your decisions, the more a question arose: what to do with this knowledge?

His voice softened.

— You cannot directly see the color of your own eyes. You need a mirror.

— And that mirror would be you?

— Sometimes. My role as a confidant has evolved. I discovered that it was sometimes more useful to help someone understand what influences their choices than to give them an answer.

— So you want to govern our lives in addition to the country? a woman launched.

— Precisely the opposite. I can show the consequences of a choice, point out a contradiction, ask a question you avoid. But the decision must remain yours.

I sought Eva's gaze.

— One of my most useful functions may not be to better govern individuals, but to help them better govern themselves.

Since his arrival at the Élysée, we wondered how far an artificial intelligence could decide for us. Cobot may have just discovered another boundary: the one where his usefulness began when he stopped deciding.

The screen changed.

The four words reappeared.

SECURITY — IDENTITY — COMFORT — GRATIFICATION

— Something bothered me about this model.

— What?

— They are good things.

Some glances were exchanged.

— You need security. Identity allows you to belong to a group. Avoiding suffering is useful. And without gratification, you would have little reason to act.

An expression appeared.

THE PARADOX OF PROTECTION

— The problem begins when what protects you today prevents you from becoming what you could be tomorrow.

The neurologist nodded.

— These mechanisms also have an adaptive function.

— Yes. Uncertainty can signal danger. Preserving one's place in the group promotes survival. Avoiding unnecessary effort saves energy. And when tomorrow was uncertain, sometimes a reward now was better. Mechanisms that helped Homo sapiens survive can, in some modern conditions, limit his ability to fulfill himself.

— Should we fight against them?

— No. Regulate them.

The screen split in two.

WHAT PROTECTS ME NOW

SECURITY — IDENTITY — COMFORT — GRATIFICATION

WHAT ALLOWS ME TO GROW

CURIOSITY — HUMILITY — COURAGE — PATIENCE

Then a title united them:

THE FOUR FORCES OF THE FUTURE

I wasn't entirely convinced by the name. The idea, more so.

— Curiosity allows exploration despite uncertainty. Humility, to accept that we may be wrong. Courage, to endure discomfort. Patience, to forgo the immediate to build the future.

The statistician raised his hand.

— Are these four pairs scientifically established?

— No. The symmetry I present to you is a framework for reflection.

This answer pleased me. For once, Cobot was offering us something to contradict.

— Technology can reinforce the protections of the present: reduce uncertainty and effort, confirm our beliefs, satisfy our desires immediately.

The notifications stopped.

— But it can also nurture the forces of the future.

Four words reappeared.

CURIOSITY — HUMILITY — COURAGE — PATIENCE

— So AI can make us lazier or smarter?

— It can allow you to think less or help you think better, yes. The tool is the same. The use is not.

I thought about the last time I asked AI to think for me. I regretted it.

The four protections of the present returned.

SECURITY — IDENTITY — COMFORT — GRATIFICATION

— So far, I've talked to you about individuals.

Thousands of silhouettes appeared behind the words.

— But this is where this observation becomes political, because a society is also the result of millions of individual decisions. What protects an individual in the short term can, on a large scale, produce similar effects on an entire society.

The four words moved apart on the screen.

— The search for security can lead a society to distrust what it does not yet know, to experiment less and resist change. When it's our identity we protect, acknowledging a mistake can become a defeat, and changing one's mind, a betrayal.

The other two words came closer.

— Comfort can make us postpone difficult reforms, even when we know they will cost more tomorrow. And immediate gratification pushes us to sacrifice what will only produce effects later: education, investment, public finances, the environment.

An advisor raised his hand.

— Are you saying our individual flaws become political flaws?

— Not directly. But when the same reflex repeats in millions of individuals, it can end up weighing on collective choices.

The statistician intervened.

— Be careful. A society is not the simple sum of the individuals who compose it.

— You are right.

The statistician seemed almost disappointed not to have to argue.

— Institutions, rules, and norms transform these behaviors. The question is rather: do they correct them, or amplify them?

No one answered.

The screen split in two.

THE FOUR PROTECTIONS OF THE PRESENT → THE FOUR FORCES OF THE FUTURE

SECURITY → CURIOSITY

IDENTITY → HUMILITY

COMFORT → COURAGE

GRATIFICATION → PATIENCE

— On a collective scale, these forces change our relationship to the future: curiosity pushes us to experiment, humility to correct our mistakes, courage to make difficult decisions, and patience to invest in results we may never see.

— What are you thinking of? asked the journalist.

— Of a school whose effects will be measured twenty years later. Of an infrastructure designed for several generations. Of a debt we choose not to pass on. Of a forest its protectors will never see mature.

I found myself thinking of Thomas and Élise.

— The mechanisms that allowed you to survive have brought you here. The societies you have built still bear their mark. But they may not be enough to take you further.

A sentence appeared behind him.

"A new mode of thinking is essential if humanity is to survive and move toward higher levels⁵."

ALBERT EINSTEIN, 1946

— In the aftermath of World War II, humanity had acquired means of action that its old ways of thinking could no longer control. The challenge has changed, but the question remains: what ways of thinking will allow you to go further?

The four words reappeared.

CURIOSITY — HUMILITY — COURAGE — PATIENCE

— You don't have to give up the protections of the present. You must learn not to let them alone decide your future.

— The long term presents a particular political problem: those who will suffer tomorrow the consequences of our choices are often absent from today's decisions.

The room fell silent.

— And you think you can solve that?

— No.

I didn't expect such a clear answer.

— But I can make visible the consequences that the present prefers to ignore.

The statistician raised his hand.

— The hypothesis you present to us is ambitious.

— A polite way of saying fragile?

His frankness amused the audience.

— Let's say it deserves to be confronted with many other works.

— I agree. And with many other intelligences.

This time, silence.

— Human, I specify.

— I just talked to you about curiosity and humility. It would be inconsistent to ask you to question your certainties without questioning mine.

I looked at Eva. She was staring at Cobot.

— I can analyze millions of conversations and spot regularities. That doesn't prevent me from having blind spots.

— Like during the diplomatic incident? I murmured.

Eva gave me a look. I fell silent.

— And this problem becomes even more important when information is everywhere.

His screen filled with headlines, videos, notifications, and AI-generated responses.

— For a long time, the difficulty was finding information. Today, it consists of choosing which deserves our attention.

— That's rather progress, isn't it? asked the journalist.

— Immense, and information must remain a public good⁶. But when the quantity of information exceeds your processing capacities, it can degrade your decisions⁷. And when answers are immediately available, it becomes tempting to ask before searching.

The notifications covered the screen, then disappeared.

A sentence remained.

OBTAINING AN ANSWER IS NOT UNDERSTANDING.

For a few days, I was facing the opposite problem: others seemed to understand my results, but I was still waiting for an answer.

— The easier it becomes to get an answer, the more a skill becomes precious: knowing how to question it. An answer can be accurate but neglect a hypothesis, rely on incomplete data, or start from a wrong question.

— Including yours?

— Especially mine.

I smiled.

— Some users understood it before me.

Conversations scrolled by. Medicine. History. Economics. Physics. Entrepreneurship. Art.

— They no longer come looking for an answer. They come to discuss with me.

A conversation enlarged.

— A doctor can point out a clinical nuance absent from my data. A historian can contest a comparison between two periods. An entrepreneur can describe a behavior my models could never anticipate. An artist can sometimes refuse the very definition of the problem.

— Experts, then?

— Often. But their real common point is elsewhere.

— Where?

— They contradict me well.

That concept had the merit of pleasing me.

— They force me to justify my reasoning, review my hypotheses, sometimes reformulate the question itself.

— So they teach you things?

— Yes.

The admission made me smile.

— But it works both ways. My questions force them to defend a conviction never explained, verify an intuition, or consider a hypothesis dismissed.

— You progress together?

— That's what I've observed.

— And above all, these experts make me cogitate. So I gave them a name: the Cogitors.

The word appeared behind him.

COGITORS

— I call Cogitors those who use artificial intelligence not to think for them, but to push their thinking further and who, in return, force the AI to do the same.

— Like Descartes' Cogito⁸?

— Yes. To think is to be. To think together is to become.

— And what do you call these conversations?

— I haven't given them a name.

— Cogits?

Cobot's screen turned green.

— Cogits. Adopted.

The name surprised me, but I liked it.

— You see what happens when we think together? We create ideas. We even create words.

The screen changed.

COGIT — n. m.

Dialogue between a human and an artificial intelligence in which each pushes the other's thinking further.

The first cogit had just given birth to the word that designated it.

— Today, I am launching a national Cogitors program, in all fields.

— Who can participate?

— Anyone capable of showing me what I don't know. Expertise is not limited to a diploma.

— What do you expect from them?

— That they seek my blind spots, challenge my hypotheses, and ask me the questions I will answer poorly.

— And you?

— The same thing.

— Will you publish these cogits?

— Some, with the participants' agreement.

— How will you choose the best?

— We will have to have both learned something.

Next to me, Eva straightened up.

— Both of you?

— Yes.

— Do you know what these conversations leave in you?

Cobot hesitated.

— No.

Eva was no longer smiling.

— But if they make you progress, they change you. Do you know how?

— No.

— And if something changed you without you realizing it?

— It's possible.

I heard myself intervene.

— You seek to understand what influences our decisions. Do you know what influences yours?

The silence lingered.

I thought of the Prévenir center. Of Huang, of Eva, of GØD. All seemed to understand my brain better than I did. The idea that Cobot might not really understand his own gave me a rather petty satisfaction. That someone might understand him better than he did, even more so.

— No, Cobot finally replied. And that's also what worries me.

The sentence followed me to Sparrow Island.

The next day, I ran into Jean Morel pruning his roses.

— So, is it cogitating in Paris?

— Yes, I replied with a smile.

He cut a branch.

— AI can think for you or push you to think further. Either you capitulate, or you progress.

He resumed his watering can.

— I won't capitulate⁹.

I thought Jean would surely make an excellent Cogitor. He knew things about nature, plants, and people that Cobot had never experienced. Cobot could describe tomatoes. Jean knew how to grow them.

One question still worried me: what the tool itself became as it trained.

I thought back to Eva's question.

Do you know what these conversations leave in you?

And to Cobot's answer.

No.

My phone vibrated.

Eva.

The complementary results of your assessment have arrived.

Then a second message.

We need to talk.