Pages

26 December 2025

Former KGB Agent, Yuri Bezmenov: The Four Stages of Ideological Subversion (1984)


The above video presents an interview with Yuri Bezmenov, a former KGB operative who defected to the West. He describes methods used to influence and destabilize societies from within. He explains a sequence of systemic shifts designed to erode societal cohesion over time, focusing on information saturation, cultural demoralization, and institutional degradation. Bezmenov discusses how weaponized narratives, media influence, and educational channels can reshape perceptions and weaken trust in governing structures. He emphasizes the role of propaganda and psychological operations in creating long-term shifts in public opinion and social norms. These concepts resonate with contemporary challenges in irregular and political warfare where adversaries use information operations, social media amplification, and subversive messaging to exploit societal divisions. Bezmenov’s breakdown reflects ongoing debates about the human domain, influence operations, and the resilience of democratic institutions.
“Yuri Bezmenov (alias Tomas Schuman), a Soviet KGB defector, explains in detail his scheme for the KGB process of subversion and takeover of target societies at a lecture in Los Angeles, 1983.
Yuri Bezmenov was a former KGB propagandist who was assigned to New Dehli, India – and defected to the West in 1970. Bezmenov explains his background, some of his training, and exactly how Soviet propaganda is spread in other countries in order to subvert their teachers, politicians, and other policy makers to a mindset receptive to the Soviet ideology. He also explains in detail the goal of Soviet propaganda as total subversion of another country and the four-step formula for achieving this goal. He recalls the details of how he escaped India, defected to the West, and settled in Montreal as an announcer for the CBC.

Pakistan’s Army Rocket Force: Strategic Leap or Burdened Gamble?

Tahir Azad 

On August 13, 2025, the Prime Minister of Pakistan announced the establishment of a new Army Rocket Force Command (ARFC). This directive would possess contemporary technology and the capability to engage the adversary from all directions. There is no revealed public information regarding the ARFC structure, size, or mission. The official statement just discusses that the focus will be on conventional missile systems rather than nuclear delivery vehicles, which remain under the prime control of the Strategic Plans Division (SPD). Some commentators thought that this announcement of an ARFC was a vital step to deter India, which is growing its missile and hypersonic capabilities. However, this ARFC has raised various questions. What is the need for raising a separate command while Pakistan already has an established strategic forces command structure? Additionally, it is also confronting many domestic challenges, such as its political instability, a suffering economy, and security problems. The discussion regarding the formation of a distinct rocket force in Pakistan, or the evolution of its current Full Spectrum Deterrence (FSD) strategy into an advanced variant known as Full Spectrum Deterrence Plus (FSD+) is pivotal to the changing geopolitical landscape of South Asia.

FSD, a concept that ensures a reaction to threats across all tiers of conflict, has long been integral to Pakistan’s deterrence strategy. Since the early 2010s, this posture has served as a robust barrier against Indian military pressure. However, the May 2025 conflict between India and Pakistan has revealed significant vulnerabilities and gaps. India utilized enhanced models of the BrahMos missile, capable of travelling at nearly supersonic speeds. These missiles successfully penetrated Pakistani defenses and struck vital targets, including those in proximity to the capital Islamabad. The strikes were alarming both symbolically and strategically, as they demonstrated that Pakistan’s air defense systems and conventional deterrent missiles were unable to consistently intercept or neutralize India’s precision-guided threats. India’s Defence Research and Development Organisation (DRDO) has successfully conducted a test of the Extended Trajectory Long Duration Hypersonic Cruise Missile (ET-LDHCM), in July 2025, a new hypersonic weapon developed under Project Vishnu. For the first time in decades, Pakistan’s authorities confront the disconcerting prospect that India might execute a limited, rapid strike campaign beneath the nuclear threshold with a significant likelihood of success.

Ambiguity as Deterrence: Why the Saudi–Pakistan Defense Pact Matters

Andrew Latham & Tani Gangal

On September 17, 2025, Saudi Arabia and Pakistan signed a Strategic Mutual Defense Agreement—an Article 5-style pledge that treats aggression against one as aggression against both. The text is terse by design, and the rollout was theater with a purpose: senior civilians and Generals put shoulder-to-shoulder to signal that this is deterrence, not diplomacy by press release. The pact’s power lies in what it makes unmistakable—collective defense—and what it leaves deliberately unclear: whether the deterrent shadow extends to a nuclear dimension.

Officials on both sides have floated and then walked back suggestive lines, tightening adversaries’ risk calculus while avoiding any explicit breach of nonproliferation red lines. No basing, transfer, or nuclear clause is disclosed; the ambiguity itself is the signal.

That ambiguity is a feature, not a flaw. Riyadh and Islamabad want a credible ceiling of risk without irreversible steps like forward-basing warheads on Saudi soil. There is no explicit nuclear provision, and Pakistan’s arsenal remains sized and postured primarily for its rivalry with India—realities that argue for disciplined signaling over dramatic announcements. Extended deterrence by insinuation complicates an adversary’s planning while preserving freedom of action for both capitals.

Pakistan in Gaza: Jihad on Steroids

Khaled Abu Toameh
Source Link

Pakistan, it seems, has stepped forward, selflessly offering to be part of US President Donald J. Trump's proposed International Stabilization Force (ISF) in the Gaza Strip. "We're very grateful to Pakistan for their offer to be part of it, or at least their offer to consider being a part of it," US Secretary of State Marco Rubio told reporters on December 20.

Under Trump's 20-point peace plan to end the Gaza war, sparked by the October 7, 2023 Hamas-led invasion of Israel, the primary goal of the ISF is to create a secure, demilitarized, and "terror-free" Gaza Strip that poses no threat to its neighbors. The ISF would accomplish this by enforcing the dismantlement of the Palestinian terror groups' military infrastructure, including tunnels and weapons-production facilities. The Trump plan, in addition, calls for establishing an "interfaith dialogue process will be established based on the values of tolerance and peaceful co-existence to try and change mindsets and narratives of Palestinians and Israelis by emphasizing the benefits that can be derived from peace."

Mahmudul Hasan Gunobi: Ascendant Leader of al-Qaeda Affiliate in Bangladesh

Iftekharul Bashar

Mahmudul Hasan Gunobi (Bangla: মাহমুদুল হাসান গুনোবি) is one of the enduring figures in Bangladesh’s Islamist militant scene. Designated by Bangladesh’s law enforcement as the spiritual leader of the al-Qaeda-affiliated Ansar al-Islam (Arabic: أنصار الإسلام, AAI), Gunobi represents the critical link between popular religious preaching and violent extremist recruitment. His acquittal and public re-emergence in 2024, following his high-profile arrest in 2021, underscores his and AAI’s resilience against government crackdowns.

Bangladeshi authorities assert that Gunobi’s importance is as a spiritual and philosophical leader. While not a military commander, he provides recruits with the theological foundation and psychological conditioning necessary to turn ordinary young men into hardened militants and suicide bombers. His operations reveal sophistication: he uses various religious platforms to mask clandestine recruitment and training networks, enabling AAI to propagate its ideology and motivate fighters to carry out lethal operations (Prothom Alo, July 14, 2021; Bangla Tribune, July 17, 2021).

The Arab Spring’s Painful Lessons

Alexander Langlois

The Arab Spring carries multiple meanings for the many millions of people across the Middle East and North Africa, let alone the world. The widespread calls for civil liberties and democracy across the region certainly were divisive, with some defining the uprisings as imperialist plots while others viewed them as a long-anticipated moment for freedom fighters and democrats who had long suffered under some of the most autocratic rulers of the 20th century. But what did this moment of national and regional upheaval truly mean for the region, its autocrats, and the people stuck under their boots, and what does that mean for the future?

On December 17, 2010, Mohamed Bouazizi’s act of self-immolation in protest against the brutal Tunisian regime of President Zine El Abidine Ben Ali unknowingly kicked off the region’s largest democracy wave since decolonization. Within a series of months, protests spread against autocratic regimes in Yemen, Egypt, Libya, Bahrain, and Syria, shocking a world long accustomed to and benefitting from repression in that part of the world. Egypt’s Hosni Mubarak regime, for example, was long viewed as a bulwark against the Arab street and the forces of Arab and Islamic nationalism that, for many in the West, posed a threat to their regional interests—namely energy flows, Israeli security, and fighting violent extremism.

The Arab Spring 15 Years Later

Seth J. Frantzman

The pro-democracy movement marked the death knell of Arab nationalism and unintentionally quickened a shift of regional power toward the Gulf States.

In early December, Tunisian authorities arrested a well-known opposition activist. Human Rights Watch noted that Ayachi Hammami, “a lawyer and rights defender, was arrested on December 2 in his home in a suburb of Tunis. Earlier that day, Hammami’s lawyers had filed an appeal before the Cassation Court, the highest court in Tunisia, and an additional request to suspend the verdict execution pending a final decision.” The crackdown on various critics and opposition elements is another step by the current leader, President Kais Saied, to cement control.

The arrests in Tunisia are an example of how one of the central countries of the Arab Spring has transitioned from a nascent democracy back to a form of authoritarianism. The Arab Spring began after Mohamed Bouazizi set himself on fire in December 2010. A month later, after protests swept the country, Tunisian President Ben Ali fled into exile. He had been in power since 1987 and had become a symbol not only of the Tunisian regime but also of Arab nationalism and secularism that had emerged in the Middle East after the colonial era.

What We Know About U.S. Interceptions of Oil Tankers in Venezuela

Genevieve Glatsky

A frame grab from a video posted on social media by Kristi Noem, the homeland security secretary, showed a helicopter flying over Centuries, another oil tanker, which was intercepted by the U.S. Coast Guard on Saturday.Credit...Agence France-Presse, via U.S. Homeland Security Secretary

President Trump’s drive to crack down on vessels moving oil from Venezuela, an escalating part of his pressure campaign against the government of Nicolás Maduro, took an unusual turn over the weekend.

In the Caribbean Sea on Saturday, the U.S. Coast Guard tried to intercept a tanker called the Bella 1, which officials said was not flying a valid national flag, making it a stateless vessel subject to boarding under international law. U.S. officials had obtained a seizure warrant for the Bella 1 based on its prior involvement in the Iranian oil trade, but officials said the ship refused to submit and sailed away.

Trump is losing sight of America’s real terrorist threat


The Trump administration has been reallocating scarce federal resources to combating drug cartels (“narco-terrorists”), the Venezuelan state (“a foreign terrorist organization”) and leftist groups like antifa (a “violent fifth column of domestic terrorists”). Aside from obvious concerns about legality, these actions also raise serious questions about the administration’s priorities and distribution of resources.

Drug cartels may be evil, but they are ultimately driven by profit and not by a murderous ideology like the Islamic State is. Antifa is a loose-knit group of activists who may be guilty of scattered acts of violence, but they’re not plotting mass casualty events like al-Qaeda does. The Venezuelan regime is complicit in human rights violations and drug trafficking, but it is not a state sponsor of terrorism like, say, Iran.

While the administration focuses on pseudo-terrorists, it risks losing focus on the battle against actual no-kidding terrorists. Just a week ago, a father-son team of ISIS-inspired terrorists killed 15 people at Sydney’s Bondi Beach during a Hanukkah celebration, shortly after an ISIS fighter in Syria killed two U.S. service members and an American civilian. (On Friday, U.S. forces bombed dozens of sites in Syria in retaliation.)

US launches operation to ‘eliminate’ ISIS fighters in Syria: Hegseth

Konstantin Toropin, Ben Finley and Aamer Madhani, 

Defense Secretary Pete Hegseth has announced the start of an operation to “eliminate ISIS fighters, infrastructure and weapons sites” in Syria following the deaths of three U.S. citizens.

“This is not the beginning of a war — it is a declaration of vengeance. The United States of America, under President Trump’s leadership, will never hesitate and never relent to defend our people,” he said Friday on social media.

Two Iowa National Guard members and a U.S. civilian interpreter were killed Dec. 13 in an attack in the Syrian desert that the Trump administration has blamed on the Islamic State group. The slain National Guard members were among hundreds of U.S. troops deployed in eastern Syria as part of a coalition fighting IS.

Soon after word of the deaths, President Donald Trump pledged “very serious retaliation” but stressed that Syria was fighting alongside U.S. troops. Trump has said Syrian President Ahmad al-Sharaa was “extremely angry and disturbed by this attack” and the shooting attack by a gunman came as the U.S. military is expanding its cooperation with Syrian security forces.

Syrian state television reported that strikes hit targets in rural areas of Deir ez-Zor and Raqqa provinces and in the Jabal al-Amour area near Palmyra. It said they targeted “weapons storage sites and headquarters

Counteroffensive Irregular Warfare: A Doctrine of Signature Reduction for Strategic Competition

Christopher Moede

Abstract: This article argues that U.S. Special Operations Forces have experienced an atrophy in counteroffensive irregular warfare capacity amid the shift to strategic competition, leaving Western interests vulnerable to adversarial gray-zone strategies. It advances signature reduction—the deliberate management of physical and digital detectability—as a human-centered doctrine capable of restoring freedom of maneuver, renewing Special Operations Forces (SOF) heritage competencies, and providing a scalable counteroffensive IW framework below the threshold of armed conflict. The author contends that institutionalizing signature reduction within IW doctrine and training is essential to preserving human primacy in an era of asymmetric technological competition.
The Strategic Imperative for Irregular Warfare

The world has gone digital. When wearable Strava fitness trackers exposed the location of previously undisclosed U.S. special operations forces operating locations in Syria in 2018, policies were quickly put in place to ban the devices. Operations continued with little risk to the mission or force. A mere four years later, when Russian surveillance equipment observed a small number of mobile devices registered in the UK on Ukrainian networks at a military base near the Polish border, 30 Russian cruise missiles tore into the facility where British volunteer fighters had been, killing 35. The technology-fueled contrast between these operational vignettes starkly exposes the hidden costs of attribution in strategic competition – tech which poses exponential risk to both mission and force beyond that which has previously been visible in the past three decades of warfare. Irregular warfare finds itself most authentically in the dynamic heart of this contrast, not as an ancillary auxiliary but rather a central character.

IISS Conflict Trends 2025


The International Institute for Strategic Studies (IISS) has published The Armed Conflict Survey 2025, which offers an in-depth analysis of the growing human and economic impacts of conflict as well as the increasing shifts in global power dynamics for the period from 1 July 2024–30 June 2025.

This interactive platform enables the exploration and comparison of data from The Armed Conflict Survey 2025 across countries and over time. Quantitative indicators complement the qualitative insights from the report, providing a comprehensive understanding of each conflict.

Data on deployments, violent events, fatalities, internally displaced persons (IDPs), refugees, third-party involvement, peace missions, humanitarian funding, criminal markets and international migrant stock by country of origin can be analysed through three interactive assets: the IISS Conflict Trends Map, Race Chart and Pie Chart.
Inclusion of a territory, country or state – or terminology or boundaries used in graphics or mapping – in these interactives does not imply legal recognition or indicate support for any government or administration.

Compass Points - Still the King


Artillery has become the decisive weapon of the war in Ukraine. With tens of millions of shells fired and drones guiding nearly every strike, this conflict has turned into one of the most intense artillery battles in modern history.

In this episode we break down how artillery systems have actually performed in Ukraine — not in theory, not in brochures, but on the battlefield. Using visually confirmed loss data, open-source intelligence, and real combat footage, we compare Soviet-era artillery, Western-supplied systems, rocket artillery, and self-propelled guns to see which platforms survive, which get destroyed, and why.

This video follows on from our previous deep dives into tanks and infantry fighting vehicles, and completes the picture of how modern combined-arms warfare is being reshaped by drones, counter-battery fire, and extreme attrition.

The World Will Regret Its Retreat From Globalization

Eswar Prasad

Globalization was meant to bring the world closer together, enmeshing advanced and developing economies in a web of mutually beneficial economic and financial linkages. From about the mid-1980s, trade and financial flows between countries expanded rapidly as governments dismantled barriers to these flows.

Not everything went according to plan. Tensions rose as the benefits were not equally shared within or among countries. Widening economic inequality, often attributed to free trade, roiled many advanced economies and has had far-reaching political consequences. While they benefited from access to foreign markets for their exports, many emerging market countries were ravaged by volatile capital flows and the fickleness of international investors. Still, there was a broad consensus that shared economic interests would ultimately triumph and even help smooth over geopolitical frictions.

Is Globalization a Lost Cause?

Keith Johnson

In The World’s Worst Bet, David Lynch sets out to explain what went wrong with globalization—an idea that was once the hope of a post-Cold War world for growth, prosperity, and peace, but instead became a vehicle for displacement, division, and polarized politics.

The problem, as Lynch, a veteran economics reporter, makes clear in a trip through more than three decades of recent history, is that so very much went wrong: the China shock, or the sudden explosion of low-cost exports beginning a quarter-century ago that shattered many working-class communities; the relentless drive for corporate efficiency and profits; increasingly vulnerable global supply chains; the increased financialization of the economy; the 2008 financial crisis; the pandemic; and the wars, especially in Eastern Europe and the Middle East. It’s little wonder globalization went astray.

Europe has lost all credibility in the Middle East. The way to regain it lies in Syria, Iraq and Lebanon

Nathalie Tocci

Ayear after the overthrow of Syria’s dictator, Bashar al-Assad, the former jihadi fighter turned Syrian president Ahmed al-Sharaa addressed the Doha Forum earlier this month, deftly parrying questions about his controversial past and outlining his country’s complex journey toward a participatory, rules-based system. As I listened, it struck me that, while Europe’s role in the Middle East has been severely damaged by its immoral stance on the Gaza war and its self-inflicted exclusion from Iran nuclear diplomacy, Europeans still have a role to play when it comes to its neighbours in the eastern Mediterranean.

Europe’s world has been turned upside down by Washington’s alignment with Moscow in the Ukraine war and the transatlantic rift as the Trump administration treats Europe as an adversary. Another dimension of this upheaval is Europe’s growing irrelevance in the Middle East. Only if Europeans accept that the past is behind them can they hope to regain a constructive independent role in the region.

Sharpening Policies to Deter China

Bishop Garrison

The U.S. is abdicating its role as a geopolitical leader, allowing China to build relationships, technology, and infrastructure that are rapidly expanding and solidifying China’s diplomatic influence and economic power tenfold. From its withdrawal from the United Nations Human Rights Council (UNHRC) and review of its participation in the UN Educational, Scientific, and Cultural Organization (UNESCO) to its reduction of engagement in a variety of international events to include the November 2025 G20 Summit, America continues its current path of international isolationism to its own detriment.

No matter the administration, the U.S. approach to China keeps missing the mark. The Biden Administration took a hardline or “hawkish” approach to China that included heavy export restrictions on cutting-edge technologies in an effort to protect American interests. When U.S. AI companies joined the movement by blocking Chinese developers from accessing their large language models, China’s response was to move forward despite sanctions and do more with less. In a short time, a Chinese startup, DeepSeek, released its own open-source LLM that rivaled U.S. AI capabilities while remaining significantly more cost effective.

Trump discovers Maduro’s Achilles’ heel

Vivian Salama

to Donald Trump, Venezuela was first all about narcotics. Now it is all about narcotics, oil, and the theft of American assets.

In the past week, Trump has added to his pressure campaign on President Nicolás Maduro by targeting the economic lifeblood of the regime: oil exports. The U.S. has seized three oil tankers in 11 days after Trump said on Truth Social that the United States was imposing a “TOTAL AND COMPLETE BLOCKADE” of tankers carrying Venezuelan oil that are subject to U.S. sanctions. The president added that the U.S. would also seek compensation for American assets that the Venezuelan government has seized, an apparent reference to past bouts of oil-industry nationalization by Caracas.

The new focus on Venezuela’s most abundant—and valuable—natural resource (the country has the largest estimated oil reserves in the world) was, in some ways, the clearest articulation yet of Trump’s ultimate aim. And some viewed the mention of a blockade as tantamount to a declaration of war, given that a blockade is recognized by international law as a belligerent act.

The Illiberal International

Nic Cheeseman, Matías Bianchi, and Jennifer Cyr

During the interwar years, support for revolutionary, anticapitalist parties by the Soviet-led Communist International laid the groundwork for the expansion of communism after World War II. Following the end of the Cold War, the U.S.-led international order promoted liberalism and democracy, albeit unevenly, enabling waves of democratic transitions worldwide. Today, political cooperation across borders is advancing autocracy. The momentum lies with a mix of authoritarian and illiberal governments, antisystem parties—typically but not only on the far right—and sympathetic private actors that are coordinating their messaging and lending each other material support.

The European Crisis: Origin and Future

George Friedman

The release of the 2025 U.S. National Security Strategy has brought into focus a fundamental tension that has been simmering since before President Donald Trump took office: the understanding between the United States and Europe that the geopolitical system that emerged from World War II was to be permanent. The National Security Strategy essentially says that this geopolitical relationship is obsolete, resulting in a sense that the U.S. has betrayed Europe. Thus is Europe’s crisis. In assuming that U.S. security guarantees were an enduring feature of global geopolitics, the Continent, as a whole, has made little effort to guarantee its own security.

U.S. guarantees were a direct byproduct of World War II. After 1945, the Soviet Union occupied and installed communist regimes in Eastern Europe. U.S. and British allies occupied Western Europe and formed a variety of democratic systems. The division left Western Europe extremely vulnerable to Soviet military action.

The decline and fall of the British Army

Ben Barry

Over the past half century, the British Army has been shaped by shifting strategic demands and repeated operational tests. Its combat capability expanded during the Cold War as planners prepared for conflict with the Soviet Union, before being proven — and strained — in Northern Ireland, the Falklands, the Gulf War, the Balkans, Sierra Leone, Afghanistan and Iraq. These campaigns reveal how Britain’s land forces adapted, and what was lost along the way.

Writing in the Spectator last month, the military historian Allan Mallinson wondered ‘how many people appreciate what a remarkably capable army we had — and how incapable that Army has become?’

In my book The Rise and Fall of the British Army, 1975-2025, I explain how the Army’s fighting power increased during the 1980s. General Sir Nigel Bagnall encouraged tactical innovation and a revolution in doctrine for armoured manoeuvre warfare. The Thatcher government’s rising defence budget funded new weapons, including Challenger tanks and Warrior infantry fighting vehicles, sustained military salaries and increased the size of the reserves in the Territorial Army.

Reframing cyber attribution

Virpratap Vikram Singh

In the last few years there has been an increase in public attribution of cyber incidents, where a government utilises a combination of technical evidence and intelligence assessments to publicly link those incidents to another state or to a state-sponsored actor. In 2025, at least seven states – including China, the Czech Republic, France and Singapore – have issued their first public attributions or made accusatory statements that were more explicit than before. In 2024, five states made public attributions for the first time, including Estonia, Palau and the Philippines.

Since 2020, IISS research has identified 188 instances in which 47 states, along with the European Union and NATO, have sought to publicly attribute cyber incidents to other states. There is significant variation in how specifically states identify the actor responsible and in the amount of evidence they choose to present. When highlighting the malicious activity, states may choose to present only a limited amount of the evidence they possess, relying on their political credibility to support their claims. To demonstrate links between identified threat actors and state sponsors – and to highlight their own investigative capabilities – they often present intelligence assessments. To trigger sanctions or criminal indictments, attributions need to be backed by substantive evidence.

Anthropic’s AI Attack Threatens the Strategy at the Foundation of Cybersecurity

Andrew J. Lohn

Chinese hackers successfully used Anthropic’s artificial intelligence (AI) models to run a hacking campaign almost entirely without human interaction, breaching “major technology companies and [foreign] government agencies,” the company reported in November. Nothing about any of the individual hacks was especially noteworthy or sophisticated. Still, the campaign marks a step change in cybersecurity.

AI-Assisted Hacking

The Chinese attackers did not just use Anthropic’s AI model to find a vulnerability in the digital armor. They built a hacking system that combined many distinct AI sessions that each worked on different aspects of the intrusion. Those separate sessions were coordinated by still other AI systems that were only occasionally directed by humans. Not only was the AI hacking system complex, but the attacks themselves were conducted in six phases, with many steps in each phase. In total, the AI system likely took dozens or more separate steps against each successful target.


The grey zone of cyber-attacks

Millie Marshall Loughran

Grace Cassy, Co-Founder of CyLon discusses the balance between social security and individual freedoms as well as the grey zone of cyber-attacks. Could you introduce yourself and your current roles? I’m Grace Cassy, an early-stage technology investor with a background that began in UK Government. I spent the first decade of my career as a diplomat, serving both overseas and in the UK – including several years as Foreign Policy Private Secretary to Tony Blair during his time as Prime Minister.

My early work was rooted in foreign policy and national security, but for the past 15 years I’ve focused on the technology sector, investing in early-stage deep tech companies. Much of my work has centred on cybersecurity, as well as adjacent fields such as AI, counter-fraud and defence. I co-founded CyLon, which began as an accelerator supporting emerging cybersecurity startups and evolved into a dedicated early-stage investment platform.

The Pentagon’s AI Problem Isn’t Algorithms, It’s Evaluation

Benjamin Jensen and Yasir Atalan

Artificial intelligence (AI) is the new arms race and the centerpiece of defense modernization efforts across multiple countries, including the United States. Yet, despite the surge in AI investments, both Silicon Valley and the Pentagon struggle to answer one simple question: How can decisionmakers know if AI actually works in the real world?

The standard approach to answer this question is an evaluation practice called benchmarking. Benchmarking is defined as “a particular combination of a dataset or sets of datasets . . . and a metric, conceptualized as representing one or more specific tasks or sets of abilities, picked up by a community of researchers as a shared framework for the comparison of method.” This practice allows the researchers to evaluate and compare AI model performance, for example, how well a large language model (LLM) answers questions about military planning. Yet, proper benchmarking studies are few and far between for national security.