# Who Frames the Crisis? From 9/11 to the Age of AI **Date de l'événement :** 09/09/2026 * Publié le 09/09/2026 ### Date 09/09/2026 ## Chapô **Twenty-five years after 9/11, crises are no longer told in the same way: the power to frame events has shifted, from governments to newsrooms, and then to platforms and those who publish the first images. Flavie Bagnol, a Franco-American media professional, spent more than twenty years working in New York newsrooms and the international media industry; a producer for CBS News in New York on September 11, 2001, she now works as a communication consultant, teaches at Sciences Po, and has contributed to UNESCO’s work on the ethics of artificial intelligence and communication. At a time when anyone can produce and circulate images within seconds, and when artificial intelligence is eroding even the evidentiary value of what we see, who still gets to decide what the story is?** ## Corps du texte September 11, 2001, New York. I was a breaking news producer for CBS News at Network News Service, a consortium in which CBS, ABC, and Fox pooled their resources. We shared cameras. We shared feeds. We shared pictures. It was efficient, and it gave us many more local stories across America. On September 11, it became something else. I was at home getting ready, with the local news on, when the first plane hit the World Trade Center. I was already on my way to the CBS Broadcast Center when the second one hit. The sky filled with smoke. Sirens everywhere, nothing else. People were rushing uptown in silence, leaving chaos and despair behind them. When I got to the newsroom, we went into breaking news mode. We needed more video. More sources. More angles. But this was not about filling airtime. We needed to understand what was happening so that we could tell America and the world what was really going on. The story was moving faster than we could process it. We knew. Everyone knew. We were inside history, reporting it and living it at the same time. We pieced reality together from fragments: local news crews, network feeds, satellite signals, eyewitnesses, police officers, firefighters, officials. Two French filmmakers, Jules and Gédéon Naudet, had been making a documentary about a rookie firefighter in Lower Manhattan when the first plane hit. They were not journalists. They were simply there. Jules Naudet was out on a routine street call with a battalion chief, camera running, when the aircraft came over. What he caught is one of only a handful of recordings of the first impact. That footage did not reach us. It stayed with them and became a CBS documentary broadcast six months later. Some of the most important visual evidence of the attack was recorded from the first hour, and the machinery we had built to gather everything didn’t know it excited. ### What Decides a Story in a Crisis Is Not Who Holds the Material, It Is Who Decides What the Story Is About We had assembled a complete newsgathering operation, and it could not deliver the single most important piece of evidence of the day. Volume was not the problem. Volume was not the answer either. What decides a story in a crisis is not who holds the material. It is who decides what the story is about. That morning, whoever could stand in front of a camera and be believed held the frame. New York City Mayor Rudy Giuliani became part of the crisis narrative itself. At a time when so much was unknown, physical presence became a form of communication. One image endures: Giuliani walking north through Lower Manhattan through clouds of ash and debris after escaping the improvised command post.  He and his commissioners had been trapped by the collapse of the South Tower. It conveyed leadership and resilience without a word. The media did not just report that image. They gave it meaning. Twelve days later, at the Yankee Stadium prayer service, Oprah Winfrey introduced Giuliani as “America’s Mayor.” The name stuck. Years later it was challenged in Grand Illusion, by Wayne Barrett and my CBS News colleague Dan Collins: Giuliani was walking the streets because his emergency command center had been inside the World Trade Center, the result, they argue, of years of failure to build a unified command system. The frame held anyway. On September 14, standing on a crumpled fire truck at Ground Zero with a bullhorn and his arm around a firefighter, President George W. Bush answered a rescue worker who had shouted that he could not hear him: “I can hear you. The rest of the world hears you. And the people who knocked these buildings down will hear all of us soon.” Not the White House, not a podium. The image, the place, and the words sent one message:  America is under attack, but together we will fight back. September 11 accelerated a transformation: the politician was no longer simply making history, but making history in front of the camera, and we were part of it. Sometimes, in a crisis, presence is stronger than information. ### Access, 1917 to 1991 For most of the twentieth century, the argument between the American government and the press was about access. Reporters wanted to get closer. Washington decided how close they could get. In April 1917, a week after the declaration of war, President Wilson created the Committee on Public Information. It was the first centralized federal wartime publicity operation. He put a journalist, George Creel, in charge. The committee produced pamphlets, press releases, magazine advertisements, films, and school campaigns, and it also worked with the post office to suppress opposing material. One agency did both jobs. Historians of the period treat the Creel committee as the reason the next administration built things differently. Allan Winkler's history of the wartime information effort describes the CPI as a precedent Washington was trying not to repeat. In the Second World War, the two jobs were separated. Roosevelt created the Office of Censorship in December 1941 and named Byron Price, an executive news editor at the Associated Press, to run it. Six months later, Executive Order 9182 created the Office of War Information, and Roosevelt named the CBS radio commentator and former New York Times reporter Elmer Davis as its director.  One office reviewed what the press produced. The other produced material of its own. Both were run by journalists. In 1942, the Office of Censorship issued a Code of Wartime Practices. It listed what editors should leave out. Troop movements. Fortifications. Weather. Casualty lists.The code was voluntary. It carried no force of law. Newsrooms followed it, nearly all of them.The same office also read international mail and cable traffic. That censorship was not voluntary. Edward R. Murrow was embedded before anyone used the word. He flew a bombing raid over Berlin in December 1943 and described it to American listeners. Walter Cronkite filed for United Press. They reported from inside the system.  In 1943, Congress shut down much of the Office of War Information's domestic work, on the grounds that it was producing political propaganda.  Vietnam changed that. Reporters had open access to the front lines. What they saw did not match what the government was saying. This time, reporters had open access to the front lines. Graphic television footage and skeptical anchor reports. By the first Gulf War, access was rationed again. Pools. Escorts. Security review before filing. Afterward, Washington bureau chiefs from fifteen news organizations wrote to Secretary of Defense Dick Cheney to protest. They commissioned a report. It concluded that the Gulf War was the most under covered major conflict in modern American history. Then came September 11. ### Iraq, 2003: nobody had to ask In 2003, the Pentagon embedded roughly 600 journalists with combat units, under a signed agreement setting out what they could report and discouraged everyone else from going in alone.  Reporters were closer to the fighting than in any war since Vietnam. But the coverage that decided how Americans understood the war had already happened, months earlier, in Washington and New York, where nobody was denied anything. Although I was working for NNS, I spent considerable time with the foreign desk at CBS News, drawn by my interest in international news. The editors would sometimes ask for my perspective. Dan Rather was the anchor. Working around him was intimidating. We were covering the lead-up to Iraq, and something was different this time. It was not about covering another attack. It was about what came after: the media falling in behind the president. I read the French press every morning. I could hear the questions being asked on the other side of the Atlantic, and from where I was standing, there was no way to get them into the story. The pressure was not direct. It was in the air. The questions that should have been asked in a story meeting were asked in the elevator instead, quietly. Being French in that newsroom meant something. As the rhetoric against France intensified, I could feel the temperature change. The jokes about “freedom fries” were not merely political. They marked dissent as un-American. The administration had already framed Iraq inside the post-9/11 national-security narrative, and doubt could be portrayed as unpatriotic. The questions coming from Paris were unusable, whatever their merit. I said very little. What happened in that building was not a failure of nerve in one junior producer. For months, the biggest newsrooms in America carried what the administration told them. Saddam Hussein had weapons of mass destruction. Officials said it, the front pages ran it, the evening broadcasts led with it, the Sunday shows debated it as settled. Then the search teams went in and found nothing. The reckoning came afterward. The New York Times published a note to its readers. The Washington Post ran its own accounting. Both landed on the same admission: they had leaned on official sources and buried the dissenting evidence deep in the paper, where it died. Poynter later called it a systemic breakdown of journalistic skepticism, sourcing and courage. Trust in the American press was above fifty percent when the war began. It has never been above fifty percent since. In 1942, the government asked, and the press said yes. In 2003, nobody had to ask. Access had never been the constraint. ### Virginia Tech, 2007: the public becomes the source On April 16, 2007, I was covering another mass tragedy, this time from the newsroom of the CBS morning program The Early Show. I was monitoring Facebook. Facebook had opened to the general public only in September 2006. It was still a college platform, designed for nothing like this. But students received a campus warning two hours after the gunman shot the first victims, and with only rumors they turned to the platform they already lived on. Within hours, a student had created a group called “I’m ok at VT.” By the following day it had thousands of members. It became a search engine: missing classmates, building evacuations, shelter information. Among the people reading those posts to gather news was Brian Stelter, a young writer who ran a media blog the networks read every morning; he would join The New York Times three months later. As we crossed paths on Facebook, I realized he was doing what I was doing. We were not waiting for press conferences. We were harvesting information from a platform built for college students. On 9/11, the infrastructure was broadcast: satellite feeds, pooled footage, official channels. By Virginia Tech, it was social. The people living the story were no longer merely witnesses. They were sources, distributors and, increasingly, participants in verification. Students supplied the material. We still decided what it meant. ### Paris, November 2015: the frame is set somewhere else By November 13, 2015, I was no longer on staff at CBS, but I was on their freelance roster. A former colleague at ABC’s London office knew I was in Paris. The phone rang. The brief from New York was to find an American victim of the attack in Paris. What the newsroom wanted was to find survivors and bring back a first-person account of the attack. To me, it felt that it was not the whole story; a nation was under attack. I went to the improvised command post near Bastille, where every outlet had set up tents, trucks, and satellite dishes within sight of one another. A few meters away, French crews were covering a different event. Their story was a country that would not be intimidated. That was not spin. Ten months earlier, in January, the French press had been attacked directly at Charlie Hebdo. It had already changed the way it covered terrorism, and it was not going to narrate this as a nation in shock. I worked with a young American producer whose assignment was to find survivors. That is how the two of us came to be standing beside an American woman just out of surgery, badly injured; her boyfriend had died in the concert hall. We were telling her story for the American public, the spin was different. The young journalists around me had no 9/11 in their bodies. For them it was history, not memory. They had been handed a frame and were executing it conscientiously, the way I had executed frames in New York. Nobody lied that week. The same event, the same images, the same access, and two national stories that barely overlapped. That is the clearest answer I have ever seen to the question in my title. What needed to be told in Paris was decided in New York, by people who were not there, using a template built in 2001. ### 2026: when reality itself cannot be trusted In 2001 we asked how to get more sources, so that we could understand what had just happened. In 2026, we ask how we know that what we are seeing is real. Consider what the Naudet footage was worth in 2001. Two brothers with a camera, in the right street at the worst possible moment. Its value was not that it was well shot. Its value was that no one could have staged it. Its authenticity came from the physical difficulty of having been there. That guarantee is gone, and I nearly fell for it myself.  Last year a video came across my Facebook feed showing the Sydney Opera House in flames. Three angles. Smoke pouring off the white sails. A man’s voice saying he could not believe what he was watching. I did what the newsroom trained me to do and went straight to the wires. Nothing on Reuters. Nothing on AP. No fire, no evacuation, no statement. The video was AI-generated, and once I knew that, I could see it: some parts were in the wrong place, and a building was missing. A fact-check from AAP, the Australian wire service, settled it. One post of that video had twelve million views. Checking takes time. False information does not. What has changed is not only the speed at which information travels, but who can produce it. Anyone with a phone and a platform can create, alter and circulate an image before journalists have established where it came from. ### The frame kept moving The last twenty-five years are usually told as a story of democratization: more feeds, more sources, more witnesses, everyone a publisher. From inside the newsrooms, the power to frame did not disperse. It moved. In 1942 it sat with the government, visibly, and the press consented in writing.  In 2003 it sat inside the newsroom, requiring no instruction and leaving no document. In 2015 it sat on a desk in New York while the event was in Paris. In 2026, it belongs to whoever posts first. Verification runs behind. It took me a few minutes to establish that the Sydney Opera House was not burning. By then the video had twelve million views.  Where did this come from, who made it, who benefits if you believe it: the questions have not changed, and they still take longer to answer than the image took to travel. The Naudet footage reached the public six months late. It is still the record of that morning. In 2001, presence carried authority. In 2026, presence is no longer proof. ### Thématique `#Démocratie` `#Géopolitique` **Licence :** `#CC-BY-ND (Attribution, Pas de modification)` **Langue :** `#Anglais` --- ### Navigation pour IA - [Index de tous les contenus](https://conference.sciencespo.fr/llms.txt) - [Plan du site (Sitemap)](https://conference.sciencespo.fr/sitemap.xml) - [Retour à l'accueil](https://conference.sciencespo.fr/)