In an eyewitness video analyzed frame by frame by The New York Times, Alex Pretti raises one hand and holds a phone in the other. Federal agents tackle him, and one appears to find and remove a gun holstered on his hip. Then, an agent shoots — and a second follows. They appear to fire nine more shots as Pretti lies on the ground.

The Trump administration has claimed Pretti was shot because of his legally carried gun — that the agents, later identified in records viewed by ProPublica as Border Patrol agent Jesus Ochoa and Customs and Border Protection (CBP) officer Raymundo Gutierrez, acted in self-defense. But the tool he was visibly holding in the seconds before he was killed is the one the Trump administration seems truly afraid of — and the one it’s fought harder to control.

The phone Pretti held — like the ones that onlookers used to record his killing and share it with the world — had a kind of power that the Trump administration has repeatedly recognized as both a threat and an instrument, depending on who’s using it.

The visual of Pretti clutching his phone moments before his death is emblematic of millions across the country clutching to digital evidence and online forums to make sense of the events taking place across the country. For those who oppose the federal government’s immigration enforcement tactics, technology — particularly in the form of phones and social media — has become one of the strongest defenses, whether used to alert others to ICE’s presence, organize actions and aid, or help those far away see what’s happening on the ground. For the Trump administration, it’s a visible thorn in its side.

The administration recognizes, and uses, the force of information technology. Official government accounts regularly share right-wing memes with authoritarian and white supremacist talking points, while cabinet secretaries and President Donald Trump quickly jump on X and Truth Social to relay their version of events. Shortly before Pretti’s killing, the administration used social media to counter video evidence of another killing on the streets of Minneapolis at the hands of a federal agent: that of 37-year-old Renee Good. In a Truth Social post, Trump claimed Good “viciously ran over the ICE Officer,” and pointed to one grainy and distant angle of the incident that he said made it “hard to believe” the agent was alive. A Times analysis of several angles of the shooting — including those that were much closer to the incident — found that the “agent was not in the path of the victim’s SUV when he fired three shots at close range.”

Officials across every administration have pushed back on negative press reports or downplayed their significance. Conservatives often point out that the Biden administration recognized the vast power of amplifying information across social media platforms — administration officials urged platforms to remove or limit the spread of medical misinformation during the covid-19 pandemic.

Still, the Trump administration has proved particularly willing to disregard obvious truth and particularly savvy at utilizing technology to shape its account of history. It’s curried favor with influencers it recognizes can be as effective, if not more so, as traditional media at spreading messages far and wide. Its rapid response memeification of policy issues fluently speaks the language of the internet. And it’s allied with or created platforms that facilitate the flow of information.

Trump learned an important lesson after his first term about how valuable it can be to control the very platforms where narratives spread. In 2020, a bystander’s video of Derek Chauvin, a white police officer, kneeling on the neck of a Black man named George Floyd for nine minutes spurred protests in cities around the country, and led to tangible — if often short-lived — change. Trump said Americans were “rightly sickened and revolted” by Floyd’s “brutal death,” but also blamed antifa and “professional anarchists” for provoking his pledge to deploy additional law enforcement to contain the demonstrations.

He also bears a more personal grudge. In 2021, Trump was kicked off of Facebook, Twitter, and YouTube over concern he’d incite further violence following the insurrection at the US Capitol. A year later, he launched his own social platform, Truth Social. By the time he was elected to a second term, Elon Musk, a Trump backer who briefly joined the administration, owned X, the recently renamed Twitter. And rather than ban TikTok when required by law last year, Trump claimed to extend the deadline for a forced sale until ByteDance finally reached an agreement with some of his closest allies.

A large part of the stated reason lawmakers from both parties voted to force a sale of TikTok was because they feared an adversarial government could control what narratives would ultimately reach American users. Some already blamed the app for creating a generational wedge in Americans’ views on Israel. It’s not surprising that in an age where tech platforms have been mostly left to their own devices to determine what voices can and can’t be heard, that policymakers would worry about what messages are able to get through — and realize the unmatched potential of owning those pipes. Before he bought Twitter, Musk waxed poetic about being a free speech absolutist, then almost immediately deplatformed journalists who pointed to public information about his private jet’s whereabouts. Musk’s actions show what’s long been true of social media platforms: Privately held businesses are not really a public square, yet they continue to shape what messages get to inform our reality.

That’s why the administration has also acted swiftly to counter, and sometimes suppress, narratives spread online by opponents. After right-wing activist Charlie Kirk was killed by a gunman at a university in Utah, some conservative lawmakers and activists pressured employers and platforms to take action against people who posted critical comments about Kirk, and US Attorney General Pam Bondi pledged to target those using “hate speech” before clarifying that “Hate speech that crosses the line into threats of violence is NOT protected by the First Amendment.” Reuters found that hundreds faced consequences from employers over their comments related to Kirk’s killing — while some celebrated or mocked his death, others simply reprinted Kirk’s own words or critiqued his political stances.

In several cases, administration officials have looked directly at platforms that amplify oppositional messages. Federal Communications Commission Chair Brendan Carr threatened broadcasters who aired comedian Jimmy Kimmel’s show after he made a joke about conservatives’ reaction to Kirk’s death. More recently, amid backlash in more liberal cities against ICE, Federal Bureau of Investigation Director Kash Patel promised to investigate Signal groups where users share information on immigration agents’ movements. And after criticism from the administration, Apple and Google both removed apps that let users report ICE sightings in public places.

As private sector businesses, tech platforms don’t have a legal obligation to make sure these kinds of apps or users’ posts remain available, but the First Amendment and Section 230 protect their choice to leave up or remove third-party content either way. While Section 230 has become a common target of Big Tech critics, proponents of the law warn that dismantling it could incentivize platforms to limit speech they consider risky that might invite legal liability.

Trump’s first administration recognized the power of technology, too. The president was notoriously a power user of what was then called Twitter. But at the time, his cabinet included more officials willing to push back on the most norm-breaking actions, and the president often ran into more checks on his power. Today, Republicans control both chambers of Congress, and many of the remaining lawmakers have tied their political fates to Trump. That’s created an environment where threats against speech have the potential to be more chilling than before.

Even so, in today’s social media landscape, it’s still possible for an event like Pretti’s death to shock people on all corners of the internet and move some of the unlikeliest online communities to speak out. Shortly after the shooting, when Department of Homeland Security Secretary Kristi Noem claimed Pretti was “brandishing” a weapon with the intention of inflicting “maximum damage,” many had already seen the videos for themselves. They chose to believe their own eyes instead.

After Pretti was killed and the first video spread across social media, people quickly looked for other angles. Protesters who stood on the frigid streets of Minneapolis, holding up their phones and pressing record, would play an important role, armed with the First Amendment and one of the most powerful tools to wield it.

Pretti’s phone didn’t protect him on the day he died. But those who captured his killing at the hands of federal agents helped the world to see what was happening in Minneapolis, and kept many of them from turning away. To an administration that has tried so hard to control the narrative, that’s as dangerous a threat as any.