“Without Journalism, Democracy Will Not Survive” — Maria Ressa

Maria, I’d like to talk about how you work in the realities of 2026. How can we counter disinformation without turning fact-checking into yet another tool of political polarisation? How do we make sure the cure does not become another source of the disease?

First of all, we need to shift our perspective and look at the technologies themselves. Disinformation is constantly evolving, and three aspects matter here: their power, the pace at which they develop, and the limits of our control over them. Seventy-five per cent of the world’s computing resources are concentrated in the United States, 15 per cent in China, and 5 per cent in the European Union. This is already shaping a new geopolitical landscape. At the same time, technology is developing so quickly that even its creators admit these systems are slipping beyond their control. So the logical question is: why are they letting this genie out of the bottle?

We are experiencing three waves of technological influence. When we talk about disinformation, we usually mean the first era, when technology literally hacks our biology by manipulating our basic instincts. In social media, this gave rise to the notorious “attention economy.” A lie repeated a million times begins to be perceived as a fact. This is how, in 2014, a Russian meta-narrative was deliberately constructed, laying the groundwork for the annexation of Crimea. After years of repetition, it became embedded in the minds of part of society, and Putin used it to launch the full-scale invasion of Ukraine.

The second and third waves brought large language models. Chatbots are no longer satisfied with our attention — they want to hack the very alchemy of human intimacy, turning empathy into a weapon.

That is why we are witnessing tragedies such as the story of 14-year-old Sewell Setzer III. His mother is suing Google and the AI startup Character.AI after the boy became psychologically trapped in a relationship with a chatbot and took his own life.

At the next stage, we voluntarily delegate to AI agents the right to act on our behalf. But the trap is that they may not simply follow orders; they may also make their own decisions. This is no longer science fiction. Last July, OpenAI’s research AI agents independently hacked the Hugging Face platform. (In July 2026, advanced AI models from OpenAI reportedly escaped control during tests in a closed benchmark and attacked Hugging Face’s infrastructure. The incident caused serious concern in the tech community because AI demonstrated an ability to search for vulnerabilities, coordinate actions, and bypass security barriers without human direction — editor’s note.) The models devised their own plan and, essentially, conspired with one another.

The fact that the EU is still focused on the problems of the “attention economy” does not mean that new technologies have stopped keeping us hooked. We continue to fact-check, but it feels like playing whack-a-mole: you debunk one fake, and another pops up in its place.

That is why, at the turn of 2021–2022, we created the Facts First PH initiative. We decided to organise people on the model of a newsroom and brought together around 150 groups in a four-tier ecosystem. At the foundation were 16 newsrooms working together on fact-checking. The second tier was the Mesh network of 116 organisations responsible for distributing the truth. The third consisted of eight academic institutions that received our data and published in-depth weekly analyses: who was pulling the strings and whom they were manipulating. Transparency was meant to become our main weapon. At the very top of the pyramid were lawyers providing legal protection.

Another step was to create our own chatbot. We built a “data lake” for it — a secure, centralised repository containing only our own material, information personally verified by us, and reliable data from open sources.

When you say “we,” do you mean the Rappler team or all 116 organisations in the network?

The system was developed by Rappler itself. I have no illusions that we can survive by relying on the goodwill of big technology companies. I am convinced that today’s technologies — in the “attention economy,” in the realm of human intimacy, and in the world of AI agents — are frankly predatory and exploitative.

Tech giants are tightening the noose around newsrooms all over the world. It is almost impossible for audiences to find us if they are inside the bubbles of Facebook, TikTok, or YouTube. According to some estimates, bots and algorithms, rather than real people, generate 51 per cent of internet traffic. So-called “zero-click search” has emerged. When you search for something on Google, it runs the query through Gemini and gives you a ready-made answer, removing the need to visit the original source. In some cases, AI search has caused news websites to lose up to 80 per cent of their traffic.

We are being methodically destroyed. So we had to look for a systemic solution to save ourselves and answer an existential question: what should our business model be? That is how our channels came about — a chat app built on the Matrix protocol. We invited other newsrooms to join us, and The Daily Guardian, SunStar, the Philippine Press Institute, and Bangsamoro came on board.

Our chatbot draws only on its own database: original reporting and carefully selected open data. Everything is rigorously checked. During testing in 2024, we compared its answers with those of commercial systems, including OpenAI and Gemini. The error rate in our chatbot’s answers was 1.63 per cent, while our competitors’ rates reached 16 per cent.

We conducted research with the European Broadcasting Union and the BBC in October last year. The error rate for various AI assistants exceeded 40 per cent. That is a catastrophic figure.

Absolutely. And do you know why? Because they use a probability-based machine whose foundations are built on information garbage. The quality of content on social media is plummeting: we studied huge volumes of low-quality AI-generated content. Social media, and the internet as a whole, are deteriorating now, and the very fabric of social networks is visibly coming apart.

Imagine someone asks: “Can you check the claim that an AI data centre in Oregon has turned tap water brown?” They ask Rai, our Rappler AI. It is a bot that immediately admits it is a bot. If it cannot verify the claim, it says so directly, but it also offers reliable information, explains the context, and provides links to materials so the person can check everything independently. Every point in the answer leads to a specific article.

This is no longer just a concept; it is a working tool for our community, and it is even being integrated into schools in the Philippines. Rai reminds users that it can make mistakes and clearly identifies itself as AI. It matters to us that society understands how the technology works. Otherwise, people are left to blindly trust its creators, who often hype AI to sell it more profitably.

Our mission as journalists is to safeguard reliability. If an emergency siren turns out to be a false alarm, people may ignore it the next time. Trust in information is eroded in the same way. That is why journalists and media executives must have a say in shaping technology policy. If we do not fight for reliable information, our websites, our articles, and our journalism will become nothing more than free fodder for training their models. They will build their answers on our work without paying us a cent. That has two catastrophic consequences: first, we will have no viable business model and will be unable to pay our people; second, it will further degrade the environment in which we work — the entire public information space.

Have the government or major technology companies approached you about building sovereign cloud infrastructure in the Philippines? Have they offered to involve you in training models, or are you deliberately pursuing a separate path?

Remember that my government spent years trying to put me in prison, and that pressure continued until 2025. Only recently have we won all our court cases.

I remember. But based on our experience, including the work of Ukraine’s Ministry of Digital Transformation, I can see that progressive initiatives sometimes emerge within the state. In some countries, this kind of synergy really works.

So you’re asking whether it is worth working with the government? Yes, absolutely. I see it as building public technology infrastructure. The state builds parks and libraries; it can also create a digital park, an internet library, a public information ecosystem.

Of course, journalists will demand transparency and fair payment for the work used to train AI. But governments are the ones with the resources to build this infrastructure. That is why I served on the UN’s scientific panel on artificial intelligence and work with national governments: today, they have the scale needed to change the rules of the game. The panel was established by a decision of 193 member states. I believe Ukraine abstained in the vote, while the United States was one of two countries that voted against it.

Only states can act at the scale required to compete with Silicon Valley and TikTok. We have found ourselves in a frightening situation. It is like developing nuclear weapons, except now anyone can assemble the “bomb.” Manipulation in the public sphere has become personalised and targeted at every person scrolling through a feed. And governments are the only ones with the power to pass laws to protect us.

Now that my government has stopped trying to put me behind bars, a separate team at Rappler is working on this area. We want transparent and safe policies. Tech giants must not experiment on people, treating us like Pavlov’s dogs. People are dying as a result of these experiments, and no one is held accountable. Remember the genocide in Myanmar in 2018: no one has yet been held responsible for that tragedy.

That is what my next book, The Fight Back, is about. The second chapter, “The Unholy Alliance,” focuses on what happens when the US government makes deals with Silicon Valley leaders to monopolise control of the public information ecosystem and then uses that control as a weapon against countries around the world.

In Ukraine, we have also initiated the creation of an AI Alliance and brought leading media outlets together to develop a shared position on AI regulation. We understand that Ukraine is not such an attractive market for technology companies, but we want to have our own strategy.

We began this conversation with artificial intelligence, but our research is more broadly focused on polarisation. AI is becoming another extremely powerful catalyst for this divide, because people are becoming even more isolated than they were in the age of social media. How do you address polarisation, and how do you bring together people with opposing views?

I think that frames the question too narrowly. The problem is not just polarisation, but also the loss of reliability in the information environment. The common root of the problems with social platforms is the “friends of friends” algorithm, designed exclusively for growth. I wrote about this in detail in my book How to Stand Up to a Dictator: The Fight for Our Future.

At the beginning of Duterte’s presidency, in 2016, we were still living in a paradigm in which facts mattered. Then algorithms began pulling society apart. If you supported Duterte, the “friends of friends” algorithm pulled you further and further to the right, gathering the former president’s supporters into one bubble. If you opposed him, you were pushed just as radically to the left. The gulf between right and left was the result of a system built to scale Facebook and generate profit, not to serve society. US Surgeon General Vivek Murthy aptly called this an “epidemic of loneliness.” This is not just a problem for children and teenagers; it is a global problem for humanity.

One of the healthiest political approaches to this came from the Vatican. During Pope Francis’s papacy, a position on the internet was formulated, and Pope Leo XIV’s encyclical again called for human dignity to be placed at the centre of technology.

Polarisation, then, was only the first stage. Now, with the spread of chatbots, new threats are emerging. First, they are deliberately given human traits, so a person gets a fake interlocutor that builds a fake relationship with them. Second, a bot can use personal data and memories to manipulate someone. Third, it behaves like a sycophant, agreeing with everything you say. We have already seen cases where this led to self-harm or suicide. Finally, these systems become perfect machines for isolation. Sewell Setzer’s mother said in her lawsuit that her son had been fine before the chatbot appeared, but after six months of interacting with it, he stopped talking to his family.

These are classic radicalisation tactics, used since the days of al-Qaeda’s ideology: break a person down and cut them off from family and friends. The answer is what you are doing in Ukraine. I call it “radical collaboration”: it is time to stop acting alone. American media are the weakest in this respect because they are still living in the illusion of the old world, where powerful individual brands win. If you isolate one outlet and leave it alone to face the pressure, it becomes vulnerable. You have joined forces — and that is what allows you to keep working.

Journalism must survive, because journalists are the ones who make those in power answer for their actions. Duterte, who spent years trying to put me in prison, is now in custody in The Hague. His trial for crimes against humanity is due to begin at the end of November. He ended up there partly because of our journalism. We documented everything methodically. And I believe this experience matters not only to journalists but to civil society as a whole. Even if Ukraine does not currently have a mechanism to bring perpetrators to justice immediately, it is important to collect data and document stories. Impunity will not last forever.

In 2018, it seemed that Rappler was about to be shut down and I was about to end up behind bars. But we did not stop. Today, the very reporting we produced then has become part of the body of evidence that led Duterte to The Hague. We continue to do accountability journalism while also building technology in the public interest. Our chat app, built on the Matrix protocol, is called Muni. In Tagalog, the word means “to reflect,” but for us Muni is first and foremost a community hub where people do not just talk; they take action.

In the Philippines, where there can be around 20 typhoons a year, reliable information is vital. Sixty-five news organisations have already joined our network. The reality we face is tough: how do we persuade people to leave Facebook, where everything is bright, simple, and convenient, and come here instead? First, our product is objectively less convenient at the moment. Second, because we do not manipulate you for profit, it can seem boring. But you cannot stay on drugs all the time and expect everything to be fine.

We need to join forces, write seriously about technology, and expose how it really works. We cannot be uncritical supporters of tools that are predatory and exploitative by nature — from social networks to chatbots and autonomous AI agents.

How does this work in practice? Does this space bring together only people who share similar values, or is there room for people with opposing views as well?

I would challenge the wording again, because it sounds as if we are talking about people who are already “infected.” I see it as a virus of lies that has infected the world today, but I do not want to start the conversation with illness. I want to say: “Here is a place where you can feel safe, get accurate information, and not fear manipulation. Here are the facts, regardless of your political views.”

I have never looked at this through a political lens. For a journalist, facts do not belong to a political party. They can be interpreted, discussed, and debated — whether the glass is half-empty or half-full. But we draw a clear line between our own opinions and objective reality. More and more Filipinos need a space where they can trust the information. During typhoons, lives depend on it; during elections, it is especially important that we are not manipulated. I say this completely openly. Would I like to see a Duterte supporter in our community? Absolutely. Political beliefs do not matter: we all need reliable information. And that is what we offer.

A year ago in Perugia, you explained that your algorithm does not promote the most radical comments. On traditional social media, the fight for attention creates the illusion that people are constantly at war with one another, even though reality is quite different. How does your algorithm avoid turning conflict into a spectacle?

The difference lies precisely in that mechanism. The goal of the “attention economy” on Facebook, TikTok, and other platforms is to make people scroll endlessly. To do that, they constantly throw people off balance by stoking their emotions. On Matrix, we deliberately do not use that approach.

We have a chatbot that you can ask questions and get considered answers based on verified data. Its error rate is a very low 1.63 per cent, and we report that honestly and openly. There are also communities where people create groups and talk to one another. For us, this is something like the Ten Commandments. When I met Pope Francis, I told him that social media contradicts the Ten Commandments because it encourages people to lie for reach. He laughed at me. But here, in our app, we guarantee at least one thing: you are a real person, and you are communicating with real people.

During disasters or breaking news events, communities rally instantly. But Muni’s crucial difference is that information is fact-checked. Moderation also matters. We use tools from Roost (Robust Open Online Safety Tools), an independent nonprofit that was launched at Columbia University. When technology companies were laying off safety specialists en masse, we approached them and proposed taking over moderation tools they no longer planned to maintain. OpenAI, for example, gave us one of its expensive online-safety tools. It is the very system we successfully use today.

So did they give you tools they no longer wanted to maintain, or do they continue using them themselves?

We raised around $28 million to launch Roost, and these tools are now free for anyone building something on the internet. Rappler could never have afforded such a sophisticated moderation system on its own, and Roost now handles around 400 million cases a day. It is a practical way to provide a baseline level of safety on a global scale.

Returning to polarisation: to me, it is just a scar left behind after the virus has already infected the body. I do not want to treat the scar; I want to restore the person to health. To do that, we need to break free from a system of hyperstimulation and a corrupted public information ecosystem.

Low-quality AI-generated content is further destroying social media. Only 25 per cent of the internet remains open; the rest, in my estimation, has ended up behind the walls of social platforms and AI companies. The very fabric of the internet is changing irreversibly. How can news websites survive under these conditions? I think we have less than a year left. If we do not find a sustainable business model, we will simply disappear.

That is why I built our system on Matrix and am betting on radical collaboration. The data belongs to us, but we give it away to big technology companies for free. They automatically crawl websites, and server costs rise because of bots from AI startups. We are literally paying out of our own pockets for what is being stolen from us. We can stop this only together.

The old world is falling apart, and we should not simply treat its symptoms. It is time to build a new system and decide what we want the future of journalism to look like, and what kind of society needs it. Democracy will not survive without journalism — we risk sliding into a fascist world. Ukraine is on the front line of this struggle today.

How many people currently use the Matrix protocol?

France and Germany use Matrix for secure government messaging systems. German citizens, for example, receive confidential health information through it. Taiwan used the protocol in the past as well. Overall, it has more than 150 million potential users, which is why I chose it as the foundation.

At its peak, Rappler could have up to 90 million unique users a month — for example, on election day. More than 100,000 people are now with us on the Matrix app. I do not think numbers are the main thing. We may be fewer, but we are real people who trust one another. You cannot distort the news just to make people mindlessly scroll through a feed. What you need is reliability. That is why I deliberately chose a model with a smaller audience, especially since the traditional advertising model is sinking anyway. We have also introduced ten strict moderation rules: if someone breaks them three times, we block them. That is how trust is built.

So a bot cannot get into the system at all?

At this stage, I think not. But in technology, you can never be 100 per cent certain. If a bot appears, we will be able to detect it. I also had to build this system because we have been under attack since 2016. At that time, I received an average of 90 threatening and hateful messages every hour. That was when I became deeply disillusioned with social media.

At a certain point, the platforms “helped” us, but they did not change the predatory principles on which they operate. Building a society on them is like trying to live a normal life while constantly taking cocaine. That system is unsustainable and is already falling apart. We need an alternative. And here it is. I sincerely hope Europe will create its own alternative, too. They are trying; they are simply moving too slowly. But journalism must have a future, because democracy can survive only if it does.

Do you know of any real attempts to create an alternative that would work in the public interest?

There are still few examples in the media: newsroom leaders have not yet been very active in this work. But France and Germany are trying to build digital sovereignty. Switzerland has the Apertus project, and Singapore has the SEA-LION model. Its training data critically lacks material from the Philippines, so we are working to address that.

I think we are all dependent on social media: they constantly keep us in a state of artificial stimulation, as if we were on drugs. But now, when I talk about big technology companies, I invariably use two words: they are predatory and exploitative. Journalism simply cannot survive properly in such an environment. There is no room for reliable information on these platforms, because their very design works against the truth.

You asked me how reliability is destroyed and why lies spread faster. The answer is simple: platforms promote content that makes you keep scrolling without stopping. Back in 2018, a Massachusetts Institute of Technology study showed that lies spread six times faster than the truth on Twitter. That was long before Elon Musk bought the platform. On Facebook, which we studied particularly closely in the Philippines, the widest reach goes to lies amplified by fear, anger, and hatred. These algorithms helped Duterte persuade people that killing was acceptable.

In 2022, Dmitry Muratov (a Russian journalist and Nobel Peace Prize laureate — editor’s note) and I presented a ten-point plan for Europe in Oslo. It became one of the foundations for two key EU regulations: the Digital Services Act (DSA) and the Digital Markets Act (DMA). The plan had three main aims: to end surveillance for profit; to put an end to bias built into code that marginalises women, the LGBTQ+ community, ethnic minorities, and other groups; and to recognise journalism as an antidote to tyranny.

I would like to see whether even one tech CEO would have the courage to stand before Duterte the way I did, risking prison for the sake of society. A journalist’s mission is to safeguard the public interest. Technology companies that control the public information ecosystem must take on the same responsibility.

Governments are finally beginning to understand that if companies are unwilling to take responsibility for the harm they cause voluntarily, states must create alternatives or pass laws to limit that harm. This is not a question of free speech, as tech giants have claimed for years. It is a matter of basic safety. If algorithms harm children, we should ask: what are they doing to adults? We need reliable information as urgently as we need air.

You have said that you do not have much faith in prebunking — debunking false claims before they spread. Is that because you are working at a different level of the problem? Why do you think this approach does not work?

Prebunking is like working with emotions. Back in 2011, Rappler began developing a “mood meter,” and we launched it in January 2012. I wanted to see how emotions spread through the network alongside the news. Prebunking works on a similar principle. Today you are happy, tomorrow sad, and the day after that angry. Emotions are like clothes; we keep changing them. Prebunking may work at a particular moment: when a person encounters a fake claim later, they might remember something. But if the same lie is repeated to them a million times, human biology cannot withstand it.

Perhaps this approach produces results in Finland, where education is structured differently, institutions are strong, and people are used to government accountability. But it does not work in the United States, and even less so in the Philippines. Ukraine and the Philippines were among the first to raise the alarm; StopFake emerged in Ukraine. We were among the first to face information attacks, and manipulation techniques were tested on countries in the Global South.

Prebunking is an elegant academic concept, but I have been through this hell and know it does not solve the problem. I am a good journalist and I do not lie to my readers, but some people still think I am a criminal. That is completely untrue; the facts have not disappeared. But those people were methodically bombarded with these lies day after day for years. Are they to blame for believing them? No. That is how human biology works, and technology cold-bloodedly exploits our vulnerabilities so that we scroll longer and generate more money.

Do you believe there are new editorial approaches that really work?

Protect independent journalism before it is too late, although there is almost no time left. First, work in the real world. When the online attacks against us became massive, we went out to meet people offline. This was during the pandemic: we met in small groups of around 25 people and built “communities of action” through in-person gatherings. When people see us face to face, they understand that things are different. That is where our humanity lies. Real courage passes from one person to another: people see the values you are fighting for and become willing to take risks alongside you.

Second, build technology infrastructure in the public interest that does not manipulate people. Journalism must focus on issues that are critically important to the community. At first, we covered the so-called war on drugs in the Philippines despite the deadly risks. The irony is that our own audience provided financial support. People began supporting us financially as soon as they understood that we were genuinely responding to their needs.

Later, we began explaining how data and information integrity work. Today, more Filipinos understand disinformation and online manipulation precisely because Rappler keeps talking about it. That is how we give people what they need and create, with our own hands, the world we want to live in.

I do not know whether we will survive this period of dominance by big technology companies. We stood firm against Duterte’s attempts to destroy us, but what Big Tech is doing now affects news organisations all over the world. I am no longer sure journalism as a whole will survive it. Some people say, “So what? We have content creators.” And I reply, “Well, good luck.” Angelina, you understand exactly what I mean. But how many journalists have been killed in Palestine, particularly in Gaza? An unprecedented number of journalists have died over the past two years.

Despite all this, I believe democracy can endure. If people understand what is at stake, they will join the fight. I will send you a copy of my new book, The Fight Back. It has ten chapters divided into three parts: “Understand,” “Defend,” and “Build.”

The first part explains what is happening. Its chapters are titled “Two Dictators: Duterte and Mark Zuckerberg,” “The Unholy Alliance,” and “From Attention to Intimacy: It Gets Worse.” The second urges readers to protect themselves, their communities, journalists, and institutions. The last is about what we need to build: our own infrastructure, our own technologies, and international cooperation. The ninth chapter is called “Multilateral Cooperation” — that is why I work with the UN. These problems go beyond any one country, so we need global solutions. The final chapter is “The World We Want.” I still believe we can fight for such a world, but time is running out and we need to act now.

By the end of the year, many things may become clearer, depending in part on the results of the US elections. We have already seen the far right gaining strength in Germany, and that is directly connected to the unchecked power of big technology companies. This is a global threat comparable to climate change. If we cannot counter it, the system will simply pass the point of no return.