Unpacking MI5’s Warning on a ‘New Era’ of Security Threats
The UK domestic intelligence agency’s annual update rightly focused on a whole-of-society response. But it only scratched the surface of the threat posed by algorithm-driven echo chambers.

By experts and staff
- Published
Farah PandithCFR ExpertSenior Fellow- Research Associate, Middle Eastern Studies
Farah Pandith is a senior fellow at CFR and the author of How We Win: Countering Violent Extremism in an Unstable World. Ezra Hess is a research associate for Middle Eastern studies at CFR.
Last week, Sir Ken McCallum, director general of MI5—the United Kingdom’s domestic security and counterintelligence agency—delivered his annual threat update. Speaking at MI5’s Thames House headquarters in London, McCallum warned of a “new era” for national security, one that required “a new response.”
“Sometimes national security can be talked about like a remote geopolitical contest, something best left to a handful of specialist government agencies, far from the concerns of everyday life,” McCallum said. “But the impacts of the threats we face at scale in 2026—arson, sabotage, cyber-attacks, coercion, violence—are deeply human.”
Farah Pandith, a senior fellow at the Council on Foreign Relations and an expert on extremism and counterterrorism, shared her thoughts about the speech—what it got right, what was missing, and where the UK and other governments should focus next to counter harmful ideologies.
What the Speech Got Right
Centering people. It was refreshing to hear the national security discussion—both the threats and the responses—focused on the human dimension. One of the goals of extremism is to break down cohesion within societies. As McCallum notes, “Those who mean us harm are trying…to divide our communities and turn us against one another.” Breakdowns in cohesion create the space in which extremism thrives, and to counter this threat we need to look at our response not just in terms of society as a whole but the individuals who make up that society.
Emphasizing resilience. The speech was right to focus on the importance of resilience, and McCallum identified three aspects of this: systemic resilience (such as critical infrastructure), societal resilience (countering efforts to distort and divide society) and technological resilience, particularly around artificial intelligence (AI). But what was especially important was the emphasis on how resilience should not be seen as an abstract concept, or something far away. Instead, it should be practiced and built upon every hour of every day, and everyone has a role in building it.
Expanding on teambuilding. McCallum talked about building a “team of teams,” meaning a whole-of-society approach. This is critical. From the private sector—particularly AI companies and social media platforms—to nongovernmental organizations, civil society, and policymakers, every sector of society has a role to play in weaving together a unified approach to address evolving national security threats including extremism and terrorism.
Recognizing the AI threat. It is not surprising that the threats posed by emerging technologies such as AI featured prominently, as they should. The power of AI in the hands of bad actors to imagine, create, destroy, and infect is in its infancy, yet the potential is already evident, particularly because of AI’s low barriers to entry—any militant group with internet access can produce highly personalized content. Last year, for example, the U.S. Department of Homeland Security highlighted [PDF] a video of a news anchor wearing a helmet and combat fatigues broadcast on a platform affiliated with the so-called Islamic State following the 2024 attack on a Russian concert venue. But the anchor was AI-generated, created by supporters of the group as part of News Harvest, an AI-generated program. With fast-evolving threats like this, creative initiatives such as MI5’s new AI Threat Assessment Centre make sense.
What Was Missing
Laying out the threat of algorithms. McCallum made passing references to “polarizing algorithms” and what they are “feeding [your] teenagers in their bedrooms.” But those brief nods understate the role that algorithms play in creating echo chambers of radicalization. The intelligence community is confronting a new era of radicalization in which algorithms, deepfakes, and social media echo chambers are fomenting extremism and will increase acts of terror and hate crimes. McCallum does state that younger people are becoming more radicalized by the internet, but he undersells the role algorithms and other features of social media play in shaping young, violent extremism.
Reckoning with the challenges of young people and radicalization. The speech referenced young people, noting that extreme right-wing terrorism “continues to skew young, driven by the internet” and even highlighting the case of a thirteen-year-old being investigated by MI5. But again, this seems to undersell the challenge. One of the clear lessons from working on the ideologies of extremists in the twenty-five years since the September 11 attacks is that governments have not done enough to understand child and adolescent minds—why and how they become radicalized—and that they must do far more to develop specific resilience models for different age groups. Addressing algorithms is just the start in terms of protecting young people. If we are going to talk about a “team of teams” approach, then we need to think bigger in how we develop and scale solutions around Generation Alpha and Generation Z vulnerabilities.
Addressing hate head on. McCallum’s speech omitted the word “hate” entirely, yet hate crimes are a growing problem in the United Kingdom—there was a 7 percent increase in hate crimes in England and Wales in the year to March 2026, with a 15 percent increase in religious hate crimes targeting Muslims and a 10 percent increase in such crimes against Jewish people. Addressing the multifaceted threat posed by hate crimes is a crucial piece of the puzzle in building a more resilient and secure society.
Because the types of extremist ideologies vary so widely (far right, far left, the so-called Islamic State, and so on) it is essential that we think big in terms of responses. Initiatives such as the AI Threat Assessment Centre are a start, but governments should think bigger still—for example, by considering what infrastructure is needed to support parents who notice changes in their children that could be tied to extremist ideologies. Building resilience also requires bespoke, local efforts rooted in people’s day-to-day lives. These efforts must be properly resourced. Governments for their part can help activate support for those on the ground through grants, partnerships, and other incentives.
The years since 9/11 have been sobering. No one imagined quite how complex, varied, and dangerous the threat landscape would be two and a half decades later. These threats are being turbocharged by AI. Ultimately, MI5 must focus on the sharp end of the spear when it comes to national security threats, and it is not possible in a single speech to cover every issue. But the speech could still have gone further in underscoring the importance of soft power and connecting the dots between what is happening in society—polarization, distrust, bigotry, racism—and the country’s vulnerabilities to terrorists or state actors.
This work represents the views solely of the author(s). The Council on Foreign Relations is an independent, nonpartisan membership organization, think tank, and publisher, and takes no institutional positions on matters of policy.