“If something happens that's really bad, maybe I'll have to just blame AI” ~ Donald Trump
Dispelling speculation that he had almost certainly died, Donald Trump returned to the public eye by revealing that US Space Command headquarters would be relocated to Alabama. Like always, the President of the United States used this announcement as an opportunity to engage in “the weave,” discussing a wide range of topics spanning through both foreign and domestic policy alike.
Barring those corners of the internet who remain unconvinced that the President, like his predecessor, is really even alive, none of this was of any surprise. Yet as Trump floated between the National Guard and the Supreme Court, from Russia-Ukraine to Israel-Gaza, the world was unexpectedly invited to witness a rare moment of self-reflection from the most powerful man in the world.
“It’s a little bit scary to be honest with you,” said Trump, reflecting on a viral AI video depicting the President’s transformation from a young child born in 1946 to the man the world knows today. The Trump administration has, of course, taken lengthy steps to reverse Biden-era regulation of the technology, publishing this summer an AI Action Plan that seeks to “innovate faster and more comprehensively than [its] competitors.” The dangers, however, of the new technological era we find ourselves in are not completely lost to the Oval Office:
“It's the kind of thing they do, and one of the problems we have with AI, it's both good and bad. If something happens really bad, just blame AI. But also they create things, you know, it works both ways. If something happens that's really bad, maybe I'll have to just blame AI.”
Trump was, ironically, speaking about a video showing items being thrown out of a second-story window of the White House that had been confirmed by the Trump administration as “routine maintenance.” This did not, however, deter the President, who claimed that the sealed, 600lb windows could in fact not be opened, and that any evidence otherwise was the product of advanced fabrication.
Even those warning of a dot-com-style bubble would find it hard to deny that AI could reshape the world. But, as even Trump has come to acknowledge, the challenges it poses to information integrity in 2025 cannot also be ignored.
Through its widespread accessibility, AI has fundamentally lowered the barrier to produce and disseminate content on a mass scale, influencing millions of individuals at the click of a button. The video that scared the President of the United States was not produced by a multi-billion technology company, but rather an individual TikTok account using a free website.
Though harmless on this occasion, less can be said for deepfakes that can now fool the most sophisticated face detectors by imitating minute facial signals indicative of a human heartbeat, or the voice cloning technology that convinced three foreign ministers, a U.S. governor and a member of Congress that they were speaking to Secretary of State Marco Rubio in July this year. In a world of filter bubbles and echo chambers dictated by algorithms, the power to even further distort social media feeds poses a serious threat to public cohesion in an era of polarisation.
Across the Atlantic, the UK has faced similar challenges. Last year, social media bots spread misleading information under the hashtag #LondonVoterFraud during the city’s mayoral election, while then-shadow health secretary Wes Streeting was the victim of two separate fake audio clips in the build-up to last July’s general election.
Research from the Alan Turing Institute found that the total number of viral AI disinformation during 2024’s UK election cycle was relatively low (16). However, this statistic fails to depict the true impact of the technology on day-to-day content consumption. While research by NewsGuard shows that Large Language Models (LLMs) are being intentionally flooded with propaganda to take advantage of consumers confirmation bias, AI is also allowing bad actors to locally target users with personalised disinformation instead of seeking mass virality.
So, what can be done? Barring a messy public breakup with Big Tech, the Trump administration is almost certain to continue its deregulatory stance towards social media and AI. Though the decisions in Washington, D.C. have meaningful consequences both in the UK and beyond, this does not mean information integrity is destined to be doomed. The Council for Countering Online Disinformation believes that, through a multi-faceted, whole-of-society approach, disinformation can still be kept at bay.
Despite American pressure, the Online Safety Act must not be watered down. Criticisms of the legislation are valid, but social media platforms cannot be made even more fertile ground for AI-generated disinformation to wreak havoc. Meanwhile, efforts to implement digital and media literacy must continue to be expanded, ensuring the British public are acutely aware of the opportunities and the threats posed by modern technology and within our information landscape. Further investment must also be directed towards AI technologies utilised for good, with detection systems and content verification tools receiving the resources they need to keep pace with increasingly sophisticated deepfakes and synthetic media.
Ultimately, the challenge is not to halt technological progress, but to ensure that as AI capabilities expand, our defences and public awareness evolve in tandem. The President's moment of candour about AI reflects a broader recognition that we stand at a crossroads. The choices made today about regulation, education, and investment in protective technologies will determine whether AI becomes a tool that strengthens our society or one that erodes the very foundation of shared truth which democracy depends upon.