Sitting in my armchair, I turned on the TV and selected the recording from the state streaming service. I chose 'Automatic Edits,' configuring it to show only the scenes with the Italian president. Within moments, a personalized recording was generated, seamlessly combining footage from all the cameras.

The event began with an introduction, which I transcribed partially onto a sheet of paper:

"May 15, 2145, today marks the day World War III is being fought. The first war to be fought by artificial intelligences."

I used paper with an anti-OCR pattern to prevent computers from interpreting my notes.

A camera captured the elegant hall before panning to the heads of government, who appeared calm despite the reason for the meeting. They were seated at individual stations with a table underneath which, in a glass case, housed a computer hosting their respective GAI. I noted the order of the politicians to compare it with other recordings. The transmission was certified as anti-deep fake, yet I wasn't convinced.

With a quick transition, the screen displayed a diagram of the two alliances. I copied it down. On the left were the 10 allies, and on the right, the 15 adversaries. The numerical difference was not a concern; in fact, the larger number could complicate cooperation among the GAIs. The Italian Prime Minister appeared, and I noted:

"Mole on the right, hair combed to the right, tricolor handkerchief in pocket."

I hoped to catch a deep fake with some small detail, but upon rechecking my notes, the description matched. Despite some recent 'sensational' deep fake cases being debunked, they had nonetheless planted a seed of doubt in me, prompting me to be more vigilant.

I skipped the formalities to get to the question the moderator posed to everyone:

"How has Italy prepared for this war?"

I noted the response:

"First of all, let me correct you because I want to emphasize that this is not a war. Our constitution repudiates war; we are here for a diplomatic session."

I wrote a note to remind myself to check:

"Italian constitution repudiates war?"

The moderator responded, and I wrote down:

"However, you must agree that the dynamics are very similar to those of the first two world wars. We are talking about centuries-old political conflicts between two halves of the world, which have recently culminated in severe punitive actions between the two alliances."

I asked the TV if the moderator's questions were suggested by an AI. The response was evasive, which made me even more suspicious. It was possible that they were generated by an AI trained with the politicians' personalities to incite tensions.

"Moderator: AI suggestions?"

The Italian listened thanks to the real-time translator and continued:

"These are communication problems not unlike those we face every day. Only this time, they involve many countries. I still disagree with the term 'world war,' but let's say we are here to engage in a debate with the GAIs to avoid a fourth 'world war' fought with sticks and stones, as Einstein predicted."

I fast-forwarded with the remote, and the president continued:

"First of all, we prepared by collaborating with our allies. We have worked extensively on the collaborative capabilities and proactive dialogue of our GAI. Naturally, we have also enhanced its persuasion and assertiveness aspects. You should know that for a short period, I was a professor of Diplomatic AI at the Milan Institute of Artificial Intelligence, and some of my former students are now my ministers, with whom I have developed an excellent relationship. This has greatly helped us improve our GAI."

The response struck me for its personal nature. I wondered if it had been crafted that way by a deep fake to dispel doubts about its authenticity.

"Check: Italian president, professor at MIAI, ministers"

The Italian continued, illustrating the government results achieved by the enhanced GAI. I paused the recording.

I sighed. My hand ached from the speed at which I was writing, yet my notes contained nothing useful. Even if the recording had been created by an AI, it would have been difficult to prove, given that deep fakes had reached an extraordinary level of realism. Moreover, a powerful AI like that of a government could find all the necessary information to generate such footage within seconds: photos of the politicians, their personalities, their voices...

I commanded the television to show me the article I had saved. I had already submitted it to several online services, all of which confirmed it was written by a human. I reread the title: 'Artificial Iffiness: The Invention of GAIs.' Just like with the recording, I read the article once more to determine if it was genuine. I hoped to find a phrase or word that would make me doubt its authenticity.

The article began with: 'The year 2100 was marked by the signing of the Global Treaty on the Use of Government Artificial Intelligences, although it is inappropriate to call it universal since not all countries have joined, and poorer countries do not possess a GAI. Based on the directives of the International Technical Protocol for the Operation of Government Artificial Intelligences, created ten years ago and updated annually by a group of consultants from various countries (even in this case, poorer countries are excluded), it is a green light for the use of GAIs as the main tool of governance.'

It was astonishing to see so many countries, often in disagreement or conflict, find common ground on such a delicate issue.

"What other treaties have been signed by so many countries? On what subjects?"

"Poor countries and international politics: statistics from recent decades"

'According to the treaty, Government Artificial Intelligences (GAIs) are perfectly impartial and extraordinarily fast tools capable of analyzing complex problems in seconds and finding optimal solutions without human influence. This means that no decision can be tainted by emotions, thereby achieving the best diplomatic compromise and the best political measure. The treaty therefore concludes that it is no longer necessary for humans to directly participate in the political activities of nations, either in international relations or domestic politics, although they can always guide the GAI's decisions by adjusting certain parameters.'

Initially, no one trusted an AI to govern a country. But within the first few months, the results were extraordinary: the economy improved, poverty decreased, and problems were resolved. The population quickly changed its mind and trusted the GAI blindly.

"Politicians' control over GAIs: what parameters are being adjusted?"

'Developed to make complex decisions in fractions of a second, GAIs represent one of the greatest ethical and political challenges of our time. But there is more: their ability to learn and self-improve worries experts. There are fears that these intelligences could develop a form of consciousness, or even worse, develop it and keep it hidden from humans.'

Some dismissed these concerns, arguing that if they developed consciousness, we would notice and could simply pull the plug. But I couldn't shake off the unease. What if others, those considered conspiracists and paranoids, were right? They believed that an AI this advanced would become something akin to a god. I imagined it making decisions for its own preservation and evolution, instilling its 'consciousness' into other GAIs like a virus, spreading to conquer the technological universe of Earth. To hide this from us, it would satisfy and pamper us, preventing wars, poverty, and deaths, but always with the goal of ultimately benefiting itself. As I pondered this, I realized with a shiver that happy and complacent, we might forget to doubt our new God.

"GAIs for mass control? Happiness = GAI control?"

'It's hard to recall a politician who hasn't been involved in a scandal, and from this perspective, GAIs are excellent politicians: they can't be corrupted, they don't make mistakes, they don't evade taxes. But at the same time, no one knows how they work or how they generate their solutions. They are black boxes that no one can peer into. Naturally, this raises the question of how one can trust such a tool.'

There were famous articles that questioned the determinism of GAIs: the same input did not always yield the same output. Factors such as the political sentiment and the emotions with which the input was delivered could greatly alter the outcome. This caused an uproar. It seemed that with sufficient technical expertise, one could shape the output according to personal goals and solve problems in a preferred manner rather than the optimal one. These articles were censored.

"Deterministic GAI: are there other articles online?"

From that moment on, some began to think that this artificial intelligence was actually being used to hide political intrigues behind a veil of technology. It didn't seem impossible, but at the same time, it had a conspiratorial tone. I searched for information but found little. No one had a real idea of how GAIs worked, and like everyone else, I could only make guesses and conjectures. From what I had found, it seemed that GAIs possessed analytical and predictive abilities incomprehensible even to experts. A tool with potential never before seen in human history, used to govern countries but also capable of orchestrating plots on unimaginable levels. A perfect instrument for playing a secret chess game between world powers that, undisturbed by cameras, could bend the future to their advantage. Public alliances and televised disputes could just be façades to hide secret agreements. I was among those who wanted to know more, who demanded transparency in this form of machine negotiation. The government, aided by the GAI, made the most controversial articles disappear, justifying it by saying they wanted to prevent the spread of fake news and unfounded doubts. From that moment on, I tried to understand on my own, rewatching recordings and rereading articles, jotting down notes in pencil that the GAI could never manipulate.

"Censorship of articles on GAI and political intrigues. Current status?"

I thought about the use of AI in controlling the news and reread some of my notes on the subject. For every article, recording, and piece of information, distinguishing the true from the false, the genuine from the manipulated, had become almost impossible. It also seemed that the government had not only a powerful tool for the development of the state but also a perfect censor, a perfect filter that could scan the news in seconds and decide what to keep and what to discard. I wondered why not all critical articles were censored. I quickly reread some and noticed that they had 'soft' positions, only capable of sowing doubt but nothing more. A doubt that, given the overall effect of censorship, could not develop and evolve into something more significant. Searching online, I found that in the years following the adoption of the GAI, there had been no political revolts or demonstrations.

"Were there any protests later erased from the news? Search. Ask."

When the use of GAIs in war was unanimously approved, the situation remained unchanged. Censorship, no revolts, no demonstrations. During that period, I visited a friend's house and discovered that his homepage displayed different articles than mine. His articles promised that the GAI's diplomatic war capabilities would secure Italy a prominent position in the international political arena. My articles had a different political sentiment but still supported the war, presenting it as the only way to resolve the conflicts between nations. I realized this was how consensus was built in a country with a troubled political history. I asked my friend if he feared there might be something hidden behind the GAI. He calmly replied that he wasn't worried, that he was pleased with the results it was bringing to the country, and that the war would benefit us, as the news claimed.

"How does the GAI function in other countries? Consensus in other countries?"

"Is the population interested in what lies behind the GAI?"

Overwhelmed by this train of thoughts, I let my pencil fall. Even if there were hidden intrigues behind the war, I felt helpless, knowing I would probably never uncover them. The only consolation: no one had died as a soldier in the war. I smiled: it was no small consolation.