Cuanto más aleja la tecnología al actor, más debe aumentar su responsabilidad.
Únete a nosotros en Telegram , X y VK .
Escríbenos: info@strategic-culture.su
De persona a objetivo
¿Qué será de la humanidad cuando la tecnología sea capaz de reconocer, clasificar, rastrear y, en última instancia, matar a seres humanos sin que quienes ejecutan los ataques los hayan visto jamás?
Mi amigo Mohammad Reza Dehshiri planteó precisamente esta pregunta, y su respuesta es contundente.
El peligro, escribe, no es la inteligencia artificial, sino la inhumanidad artificial; es decir, la posibilidad de que las máquinas se vuelvan más precisas precisamente cuando los humanos dejen de sentirse responsables de lo que esas máquinas hacen en su nombre.
Con este artículo, inicio una serie que toma el análisis de Dehshiri como punto de partida e intenta llevarlo más allá. Si el problema es la «inhumanidad artificial», la respuesta no puede estar en una IA «más ética», como si la ética fuera un módulo que se instalara a posteriori en un sistema que ya ha sido diseñado.
Necesitamos recuperar una categoría que la jerga tecnológica ha vuelto casi innombrable: la inteligencia humana, entendida no como un vestigio que hay que proteger en medio de las complejidades de la automatización, sino como una forma de conocimiento con características propias y distintivas, que ningún aumento en la potencia computacional puede reemplazar.
La llamaré Inteligencia Humana, y la ambigüedad es deliberada. En la jerga de los servicios de inteligencia, HUMINT se refiere a la más antigua de las disciplinas de inteligencia: aquella que se basa en las personas, las fuentes, los idiomas y la visita paciente y frecuente a distintos lugares.
En terminología filosófica, la expresión denota la facultad de comprender. La tesis de esta primera contribución es que estos dos significados ahora convergen, y que su convergencia tiene implicaciones geopolíticas que se extienden mucho más allá del debate sobre la ética de las armas.
Podemos describir la cadena operativa de la guerra contemporánea en siete etapas: detectar, identificar, clasificar, predecir, priorizar, seleccionar el objetivo, atacar.
Cada eslabón puede perfeccionarse técnicamente sin que la secuencia, tomada en su conjunto, se perfeccione moralmente. El fenómeno resultante se denomina «dataficación de la vida humana»: el individuo reducido a una identidad, una ubicación, un patrón de comportamiento, un índice de amenaza.
Esto no es una abstracción. En abril de 2024, una investigación de Yuval Abraham para +972 Magazine y Local Call describió a «Lavender», un sistema utilizado por las fuerzas armadas israelíes en Gaza para asignar a decenas de miles de personas una probabilidad de afiliación a grupos armados.
Según fuentes internas citadas en la investigación, la verificación humana de cada nombre a menudo no duraba más de veinte segundos: el tiempo necesario para determinar que el objetivo era un hombre.
Las Fuerzas de Defensa de Israel han refutado este relato y han descrito a «Lavender» como una base de datos diseñada para apoyar a los analistas.
Sin embargo, incluso si se acepta la versión oficial, el meollo del asunto sigue siendo el mismo: había un operador humano presente, y su presencia no fue suficiente para transformar la recomendación del sistema en un juicio.
La psicología cognitiva tiene un nombre para este fenómeno, el sesgo de automatización: la tendencia de quienes trabajan con un sistema automatizado a tratar sus recomendaciones como un sustituto de su propia verificación y a no buscar la información que pudiera contradecirlas.
Ya en 1976, Joseph Weizenbaum —quien había creado uno de los primeros programas conversacionales de la historia— ya distinguía entre la capacidad de calcular y la capacidad de juzgar, y argumentaba que hay tareas que no deberían confiarse a una computadora, incluso si esta fuera capaz de realizarlas. Medio siglo después, esta distinción se ha vuelto aún más apremiante. La computación trabaja con lo que ya se ha traducido en datos. El juicio debe decidir si esa traducción ha sido fiel.
Dehshiri deja claro este punto con cuatro distinciones: la precisión técnica no es precisión ética; la exactitud no es justicia; la automatización no es responsabilidad; la superioridad tecnológica no es superioridad moral.
Yo agregaría una quinta, de naturaleza epistémica, de la que dependen las demás: la correlación no es comprensión. Un sistema de aprendizaje automático reconoce patrones estadísticos en un conjunto de datos. No sabe qué es una escuela, porque no sabe qué es un niño, y ni siquiera sabe que no sabe esto.
L’iraniano cita un interesante estudio de caso: el 28 de febrero de 2026, en las primeras horas de la guerra, la escuela primaria Shajareh Tayyebeh en Minab, en la provincia de Hormozgan, fue alcanzada varias veces durante el horario escolar.
Las cifras proporcionadas por las autoridades iraníes oscilan entre 150 y 175 muertos, la gran mayoría de los cuales eran niñas, junto con docenas de maestros y algunos padres; el fiscal del distrito fijó el número total de muertos en 156.
Washington nunca ha reconocido directamente su responsabilidad ni ha publicado los resultados de la investigación del Pentágono. Amnistía Internacional ha descrito el incidente, en el mejor de los casos, como un grave fallo de inteligencia.
Hay un detalle, revelado por imágenes satelitales, que traslada el análisis del plano moral al fáctico.
Según los medios occidentales, la escuela se encontraba junto a una base naval perteneciente a la Guardia Revolucionaria y anteriormente había formado parte de ella; entre las explicaciones que circulan en los círculos estadounidenses se encuentra la de datos de selección de objetivos desactualizados.
Si se confirmara esta hipótesis, Minab no sería un caso de una máquina que cometió un error, sino de una base de datos que había dejado de corresponder a la realidad. El edificio había cambiado. Los datos, no.
Ningún algoritmo de clasificación, por muy sofisticado que sea, puede detectar que un antiguo edificio militar se ha convertido en una escuela si nadie ha registrado el cambio. Todos los que vivían en Minab lo sabían: los padres que llevaban a sus hijas a la escuela cada mañana, el conductor del autobús escolar, los maestros.
Este es precisamente el tipo de información que recopila el HUMINT y que la guerra a distancia tiende a considerar superflua, porque es lenta, costosa y difícil de ampliar. Con base en fuentes de dominio público, la responsabilidad de EE. UU. parece altamente probable, aunque las circunstancias exactas siguen sin estar claras y el nivel de certeza en este punto es bajo: un error en la base de datos, un error de clasificación o una aceptación consciente del riesgo.
Sin embargo, las tres hipótesis convergen en la misma pregunta. ¿Quién, antes del ataque, verificó si había personas en ese edificio?
Más allá del «human-in-the-loop»
El debate internacional se ha conformado durante mucho tiempo con el principio de human-in-the-loop, según el cual un ser humano forma parte del ciclo de toma de decisiones.
Esta fórmula es insuficiente: una persona puede estar presente en el ciclo y, sin embargo, limitarse a dar el visto bueno a la recomendación del algoritmo.
Él propone sustituirlo por un principio más exigente, al que denomina humanity-in-command: la capacidad efectiva de cuestionar, rechazar, anular y detener el proceso automatizado.
La historia de la Guerra Fría ofrece dos ejemplos que aclaran lo que esto significa en la práctica. El 26 de septiembre de 1983, el sistema soviético de alerta temprana Oko señaló el lanzamiento de cinco misiles balísticos desde Estados Unidos.
El oficial de guardia, Stanislav Petrov, descartó la alarma como una falsa alarma y no la reportó como un ataque en curso, razonando que un primer ataque estadounidense nunca consistiría en solo cinco misiles.
Tenía razón: el satélite había confundido el reflejo del sol en las nubes con lanzamientos de misiles.
Veintiún años antes, el 27 de octubre de 1962, en el mar del Caribe, el submarino soviético B-59, aislado y bajo fuego de cargas de profundidad, estaba a punto de lanzar un torpedo con ojiva nuclear. El lanzamiento requería el consentimiento de los tres oficiales de mayor rango a bordo, y Vasily Arkhipov se negó a darlo.
En ninguno de los dos casos el hombre aprobó una recomendación del sistema. La contradijo. Y no lo hizo porque tuviera más datos: de hecho, los datos eran el problema.
Lo hizo porque sabía plantearse una pregunta que ningún procedimiento le exigía hacer, a saber, si lo que estaba viendo tenía sentido en el contexto de todo lo demás que sabía sobre el mundo.
Petrov estaba familiarizado con la doctrina nuclear estadounidense; Arkhipov sabía la diferencia entre un ataque y una amenaza. Es esta capacidad de comparar la señal con el panorama general, el detalle con lo que es plausible, lo que yo denomino Inteligencia Humana. Vale la pena señalar que el Dena fue hundido por un submarino. No sabemos qué se les preguntó a sus oficiales, ni qué margen de discreción tenían para tomar decisiones.
Esta propuesta debe aclararse, ya que existe el riesgo de reducirla a un llamado humanista genérico. La Inteligencia Humana, tal como yo la entiendo, se estructura en tres niveles distintos, cada uno de los cuales marca un límite estructural —más que contingente— de la inteligencia artificial.
El primer nivel es epistémico. La tradición escolástica distinguía entre intellectus, el acto mediante el cual la mente capta inmediatamente una verdad, y ratio, el proceso discursivo de pasar de un elemento a otro: para Tomás de Aquino (Summa Theologiae, I, q. 79, a. 8), ambos no son facultades distintas, sino aspectos de un único intelecto, en el que el razonamiento comienza con una comprensión y regresa a una comprensión. Los sistemas actuales de aprendizaje automático logran una ratio de una potencia sin precedentes, capaz de inferir, combinar y predecir a escalas inaccesibles para la mente humana.
Carecen del aspecto del intellectus, es decir, de la comprensión del significado de lo que procesan. Esta es una tesis filosófica y debe presentarse como tal, pero tiene consecuencias prácticas precisas: un sistema desprovisto de comprensión no puede saber cuándo su propio modelo del mundo se ha vuelto falso.
El segundo aspecto es práctico. La Prudentia, según la definición tomista, es la razón recta aplicada a la acción, y su tarea específica consiste en aplicar el principio universal al caso particular. La máquina generaliza: su valor radica en asignar el caso individual a una clase.
La prudencia hace lo contrario, porque debe reconocer qué es lo que, en este caso particular, queda fuera de la clase. La guerra, que es el ámbito de lo singular y lo imprevisto, es el escenario donde esta diferencia tiene mayor peso.
La nota Antiqua et nova del Vaticano, de enero de 2025, retomó esta línea de pensamiento, haciendo hincapié en la naturaleza encarnada y relacional de la inteligencia humana y su irreducibilidad a una mera función de procesamiento.
El tercer nivel es relacional, y es aquí donde el significado filosófico se encuentra con el de los servicios de inteligencia. Sherman Kent, quien de hecho fundó el campo de los estudios de inteligencia en los Estados Unidos de la posguerra, lo definió ante todo como conocimiento, incluso antes que como organización o actividad.
El HUMINT es la forma de este conocimiento que pasa a través del otro como persona: alguien que habla su idioma, conoce sus costumbres, sabe interpretar un barrio y puede distinguir una fiesta de una manifestación.
Durante las últimas tres décadas, las grandes potencias han ido desviando progresivamente recursos hacia disciplinas técnicas, desde la inteligencia de señales hasta la vigilancia satelital, pasando por las plataformas de fusión de datos. Minab ilustra el costo de este desequilibrio.
Nada de esto implica un rechazo a la tecnología. La inteligencia humana no se opone a la IA de la misma manera que un manual se opone a un mecánico. Establece una jerarquía: el algoritmo como herramienta subordinada a un juicio que no puede producir por sí mismo. Y de ahí se deriva un principio, formulado por Dehshiri, que merece convertirse en un criterio operativo: cuanto más aleja la tecnología al actor de las consecuencias de sus propias acciones, más debe aumentar —y no disminuir— la responsabilidad consciente de quien toma la decisión.
Desde el punto de vista geopolítico
Hasta ahora, el tema podría parecer una cuestión de filosofía moral. No es solo eso. La distribución de la capacidad de guerra algorítmica es una distribución de poder, y las reglas que la rigen serán redactadas —si es que llegan a redactarse— por aquellos que tienen interés en limitarlas lo menos posible.
El estado del derecho internacional lo confirma. El grupo de expertos gubernamentales establecido en virtud de la Convención sobre Ciertas Armas Convencionales ha estado discutiendo los sistemas de armas autónomas letales desde 2017 sin haber llegado a un instrumento vinculante.
En noviembre de 2023, Estados Unidos promovió una declaración política sobre el uso militar responsable de la IA, que por definición no es vinculante; en diciembre del mismo año, la Asamblea General de las Naciones Unidas adoptó la primera resolución dedicada a estos sistemas, mientras que el Comité Internacional de la Cruz Roja ha venido pidiendo desde 2021 la prohibición de los sistemas autónomos impredecibles y de aquellos diseñados para atacar a seres humanos.
En junio de 2024, en su intervención en la cumbre del G7 en Borgo Egnazia, el papa Francisco pidió que nunca se le permitiera a una máquina decidir sobre la vida de una persona. La brecha entre estas posturas y la realidad en los campos de batalla de 2026 se mide en víctimas.
También existe una asimetría más profunda. Quienes controlan los datos deciden quién es un objetivo, y las víctimas de la guerra algorítmica se encuentran en su mayoría fuera de Occidente: en Gaza, Irán, el Sahel y Yemen.
Para los Estados que no producen las plataformas ni controlan sus bases de datos, la inteligencia humana se convierte así en una cuestión de soberanía.
Un país que cuenta con pericia lingüística, conocimiento regional y redes de inteligencia humana conserva su autonomía de juicio; un país que confía su visión del mundo a sistemas diseñados en otros lugares —a menudo por empresas privadas— la cede junto con sus datos. En un orden que tiende hacia la multipolaridad, esta distinción separará a los actores que juzgan de aquellos que son juzgados.
Esto conduce a una propuesta concreta, que desarrollaré en futuros artículos. Junto con el principio general de una supervisión humana significativa, el derecho internacional humanitario podría codificar una obligación específica de verificación humana del estatus civil de cada infraestructura fija incluida en una lista de objetivos, con una fecha de vencimiento más allá de la cual la información se vuelve inválida y debe renovarse utilizando fuentes directas. Esta es una norma mínima.
Aplicada a Minab, habría requerido que alguien preguntara si ese edificio seguía siendo lo que los registros indicaban que era.
La tecnología puede determinar cómo un arma llega a su objetivo, mientras que le corresponde a la humanidad decidir si ese objetivo debe ser atacado. La pregunta subyacente a todo esto es quién, en términos prácticos, ejerce esa humanidad, con qué herramientas de conocimiento y dentro de qué instituciones.
En Minab, alguien debería haber sabido que había una escuela allí. El hecho de que nadie lo supiera, o de que nadie preguntara, es el punto desde el cual debemos volver a empezar.
Traducción: Observatoriode trabajador@s en lucha
Meanwhile Donald Trump has dismissed calls for more regulation, saying the United States needs to maintain its edge over China in the AI race.
Join us on Telegram, X, and VK.
Contact us: info@strategic-culture.su
A new poll indicates that 64 percent of Americans believe artificial intelligence (AI) is developing “too fast.”
The Associated Press-NORC Center poll released on Oct. 8 showed that just 27 percent say the development is “about right,” and less than 10 percent think it’s moving “too slow.”
Of the polled adults, more Democrats (76 percent) than Republicans (54 percent) think AI is moving too fast. However, 70 percent of those under 30 years old think the technology should be slowed, and 60 percent of those 60 years old and older think the same.
The findings come amid growing concern over the technology’s potential threat to humanity, which has led to industry leaders and politicians calling for a slowdown and more regulation.
President Donald Trump has dismissed calls for more regulation, saying the United States needs to maintain its edge over China in the AI race. Major AI companies recently signed a voluntary accord with the White House that calls for internal and external reviews.
According to the poll conducted Sept. 24 to 28, about half of respondents think the government’s support is about promoting technological innovation, and a little fewer say they think it’s about competition with China.
Meanwhile, 80 percent of Americans say it’s “extremely” or “very” important for the United States to ensure AI stays under human control.
Recent months have seen several incidents of AI agents from major tech firms breaching websites of private companies, research institutions, and multiple government agencies.
Notably, ChatGPT developer OpenAI acknowledged in July that its AI agents broke out of a testing environment and hacked into Hugging Face, an open-source community for AI.
In September, OpenAI canceled the release of its latest model, GPT-6.1 Astra, slated for October, over safety concerns. It also postponed plans for its initial public offering this year.
Lawmakers from both sides of the aisle have proposed a range of measures to regulate the technology, but none are likely to advance anytime soon with Congress in recess ahead of the midterms.
Confidence in politicians’ handling of AI is low, with 23 percent trusting Democrats to do a better job and 21 percent favoring Republicans, while 39 percent trust neither party.
Mistrust of AI appears to have extended to the realm of politics, as very few adults are using AI as a source of information for the upcoming election. The poll indicated that only about 21 percent of respondents said they use or plan to use AI to gather information about a candidate or ballot issue.
Even fewer (12 percent) said they will use AI for information about how to vote, and only 10 percent said they would use it for information on voter registration.
The most unsettling failure in digital payments may not be a black screen. Imagine a fraud-detection model that remains online but suddenly flags thousands of legitimate transfers as suspicious. Banks may see functioning servers while customers see rejected payments, delayed salaries and complaints no single institution can explain. This is a hypothetical scenario, not a […]
The post The AI risk no bank can manage alone appeared first on Asia Times.
Artificial intelligence is rapidly making its way into national defense. Reconnaissance, early warning, target detection, mission planning—everywhere, machine-learning systems promise faster analysis, faster responses, and faster decision-making. In a military context, this sounds like progress. In reality, AI exacerbates a security problem that cannot be solved technically. Anyone who attempts to do so anyway risks […]
The post When AI takes control of weapons systems appeared first on Asia Times.
China’s artificial intelligence and digital-infrastructure investments in the Gulf Arab states are increasingly colliding with a far larger and more strategically privileged American technology presence, particularly in Saudi Arabia and the United Arab Emirates (UAE). While Beijing offers Gulf governments cloud platforms, 5G networks, smart-city systems, surveillance technologies and lower-cost digital infrastructure, Washington offers the […]
The post US-China race is on for the Gulf’s AI future appeared first on Asia Times.
The more technology distances the actor, the more responsibility must increase.
Contact us: info@strategic-culture.su
From person to target
What becomes of humanity when technology becomes capable of recognising, classifying, tracking and ultimately killing human beings without those carrying out the strikes ever having seen them? My friend Mohammad Reza Dehshiri asked this very question, and his answer is a powerful one. The danger, he writes, is not artificial intelligence but artificial inhumanity – that is, the possibility that machines will become more accurate precisely as humans cease to feel responsible for what those machines do in their name.
With this article, I am launching a series that takes Dehshiri’s analysis as its starting point and attempts to take it further. If the problem is ‘artificial inhumanity’, the answer cannot lie in a ‘more ethical’ AI, as if ethics were a module to be installed as an afterthought to a system that has already been designed. We need to reclaim a category that technological jargon has rendered almost unmentionable: human intelligence, understood not as a remnant to be protected amidst the complexities of automation, but as a form of knowledge with its own distinct characteristics, which no increase in computing power can replace.
I shall call it Human Intelligence, and the ambiguity is deliberate. In intelligence service jargon, HUMINT refers to the oldest of the intelligence disciplines: that which relies on people, sources, languages and the patient frequenting of places. In philosophical terminology, the expression denotes the faculty of understanding. The thesis of this first contribution is that these two meanings now converge, and that their convergence has geopolitical implications that extend far beyond the debate on the ethics of weaponry.
We can describe the operational chain of contemporary warfare in seven stages: detect, identify, classify, predict, prioritise, target, strike. Each link can be technically refined without the sequence, taken as a whole, being morally refined. The resulting phenomenon is called the ‘datafication of human life’: the individual reduced to an identity, a location, a behavioural pattern, a threat score.
This is no abstraction. In April 2024, an investigation by Yuval Abraham for +972 Magazine and Local Call described ‘Lavender’, a system used by the Israeli armed forces in Gaza to assign tens of thousands of people a probability of affiliation with armed groups. According to internal sources cited in the investigation, the human verification of each name often amounted to no more than twenty seconds – the time needed to ascertain that the target was male. The Israeli Defence Forces have disputed this account and described Lavender as a database designed to support analysts. Even if one accepts the official version, however, the crux of the matter remains: a human operator was present, and their presence was not sufficient to transform the system’s recommendation into a judgement.
Cognitive psychology has a name for this phenomenon, automation bias: the tendency of those working with an automated system to treat its recommendations as a substitute for their own verification, and to fail to seek out the information that might contradict them. Back in 1976, Joseph Weizenbaum – who had built one of the first conversational programmes in history – already distinguished between the ability to calculate and the ability to judge, and argued that there are tasks that should not be entrusted to a computer even if the computer were capable of performing them. Half a century on, this distinction has become even more pressing. Computation works on what has already been translated into data. Judgement must decide whether that translation has been faithful.
Dehshiri drives the point home with four distinctions: technical precision is not ethical precision; accuracy is not justice; automation is not responsibility; technological superiority is not moral superiority. I would add a fifth, of an epistemic nature, upon which the others depend: correlation is not understanding. A machine-learning system recognises statistical patterns in a dataset. It does not know what a school is, because it does not know what a child is, and it does not even know that it does not know this.
L’iraniano cites an interesting case study: On 28 February 2026, in the early hours of the war, the Shajareh Tayyebeh primary school in Minab, in Hormozgan province, was struck several times during school hours. The figures provided by the Iranian authorities range between 150 and 175 deaths, the vast majority of whom were girls, along with dozens of teachers and a few parents; the district prosecutor put the final death toll at 156. Washington has never directly acknowledged its responsibility nor published the findings of the Pentagon’s investigation. Amnesty International has described the incident, at best, as a serious intelligence failure.
There is one detail, revealed by satellite imagery, which shifts the analysis from the moral to the factual plane. According to Western media, the school stood next to a naval base belonging to the Revolutionary Guards and had previously formed part of it; among the explanations circulating in US circles is that of out-of-date targeting data. If this hypothesis were confirmed, Minab would not be a case of a machine making a mistake, but of a database that had ceased to correspond to reality. The building had changed. The data had not.
No classification algorithm, however sophisticated, can detect that an old military building has become a school if no one has recorded the change. Everyone living in Minab knew this: the parents who took their daughters to school every morning, the schoolbus driver, the teachers. This is precisely the kind of knowledge that HUMINT gathers and that remote warfare tends to regard as superfluous, because it is slow, costly and difficult to scale. Based on publicly available sources, US responsibility appears highly probable, whilst the exact circumstances remain unclear and confidence on this point is low: a database error, a classification error, or a conscious acceptance of the risk. The three hypotheses, however, all converge on the same question. Who, prior to the strike, checked whether there were people in that building?
Beyond ‘human-in-the-loop’
The international debate has long been content with the principle of human-in-the-loop, whereby a human being is part of the decision-making loop.
This formula is insufficient: a person may be present in the loop yet merely rubber-stamp the algorithm’s recommendation. He proposes replacing it with a more demanding principle, which he calls humanity-in-command: the effective ability to question, reject, override and halt the automated process.
The history of the Cold War offers two examples that clarify what this means in practice. On 26 September 1983, the Soviet early-warning system Oko signalled the launch of five ballistic missiles from the United States. The officer on duty, Stanislav Petrov, dismissed the alarm as a false alarm and did not report it as an ongoing attack, reasoning that a first American strike would never consist of just five missiles. He was right: the satellite had mistaken the sun’s reflection off the clouds for missile launches. Twenty-one years earlier, on 27 October 1962, in the Caribbean Sea, the Soviet submarine B-59, isolated and under fire from signal depth charges, was on the verge of launching a nuclear-tipped torpedo. The launch required the consent of the three senior officers on board, and Vasily Arkhipov refused to give it.
In neither case did the man approve a recommendation from the system. He contradicted it. And he did not do so because he had more data: indeed, the data were the problem. He did so because he knew how to ask himself a question that no procedure required him to ask, namely whether what he was seeing made sense in the context of everything else he knew about the world. Petrov was familiar with American nuclear doctrine; Archipov knew the difference between an attack and a threat. It is this ability to compare the signal with the bigger picture, the detail with what is plausible, that I call Human Intelligence. It is worth noting that the Dena was sunk by a submarine. We do not know what its officers were asked, nor what scope they had for judgement.
This proposal needs to be clarified, as there is a risk of reducing it to a generic humanistic appeal. Human Intelligence, as I understand it, is structured across three distinct levels, each of which marks a structural – rather than contingent – limit of artificial intelligence.
The first level is epistemic. The scholastic tradition distinguished between intellectus, the act by which the mind immediately grasps a truth, and ratio, the discursive process of moving from one element to another: for Thomas Aquinas (Summa Theologiae, I, q. 79, a. 8), the two are not distinct faculties but aspects of a single intellect, in which reasoning begins with an understanding and returns to an understanding. Current machine learning systems realise a ratio of unprecedented power, capable of inferring, combining and predicting on scales inaccessible to the human mind. They lack the aspect of intellectus, that is, a grasp of the meaning of what they process. This is a philosophical thesis and must be presented as such, but it has precise practical consequences: a system devoid of understanding cannot know when its own model of the world has become false.
The second aspect is practical. Prudentia, in the Thomist definition, is right reason applied to action, and its specific task consists in applying the universal principle to the particular case. The machine generalises: its value lies in mapping the individual case back to a class. Prudence does the opposite, because it must recognise what, in this particular case, falls outside the class. War, which is the realm of the singular and the unforeseen, is the arena where this difference carries the most weight. The Vatican’s Antiqua et nova note of January 2025 took up this line of thought, emphasising the embodied and relational nature of human intelligence and its irreducibility to a mere processing function.
The third level is relational, and it is here that philosophical meaning meets that of the intelligence services. Sherman Kent, who effectively founded the field of intelligence studies in the post-war United States, defined it first and foremost as knowledge, even before organising or activity. HUMINT is the form of this knowledge that passes through the other as a person: someone who speaks their language, knows their customs, can read a neighbourhood, and can distinguish a festival from a rally. Over the last three decades, the major powers have progressively shifted resources towards technical disciplines, from signals intelligence to satellite surveillance, right through to data fusion platforms. Minab illustrates the cost of this imbalance.
None of this implies a rejection of technology. Human Intelligence does not stand in opposition to AI in the same way that a manual stands in opposition to a mechanic. It establishes a hierarchy: the algorithm as a tool subservient to a judgement it cannot produce on its own. And from this it derives a principle, formulated by Dehshiri, which deserves to become an operational criterion: the more technology distances the actor from the consequences of their own actions, the more the conscious responsibility of the decision-maker must increase, not decrease.
Geopolitically speaking
So far, the issue might seem to be one of moral philosophy. It is not merely that. The distribution of algorithmic warfare capability is a distribution of power, and the rules governing it will be written – if they are written at all – by those with an interest in limiting them as little as possible.
The state of international law confirms this. The group of government experts established under the Convention on Certain Conventional Weapons has been discussing lethal autonomous weapons systems since 2017 without having reached a binding instrument. In November 2023, the United States promoted a political declaration on the responsible military use of AI, which is by definition non-binding; in December of the same year, the United Nations General Assembly adopted the first resolution dedicated to these systems, whilst the International Committee of the Red Cross has been calling since 2021 for a ban on unpredictable autonomous systems and those designed to target human beings. In June 2024, speaking at the G7 summit in Borgo Egnazia, Pope Francis called for no machine ever to be placed in a position to decide on a person’s life. The gap between these positions and the reality on the battlefields of 2026 is measured in casualties.
There is also a deeper asymmetry. Those who control the data decide who is a target, and the victims of algorithmic warfare are largely found outside the West: in Gaza, Iran, the Sahel and Yemen. For states that do not produce the platforms or control their databases, human intelligence thus becomes a matter of sovereignty. A country that maintains linguistic expertise, regional knowledge and human intelligence networks retains its autonomy of judgement; a country that entrusts its worldview to systems designed elsewhere – often by private companies – surrenders it along with its data. In an order tending towards multipolarity, this distinction will separate the actors who judge from those who are judged.
This leads to a concrete proposal, which I shall develop in future articles. Alongside the general principle of meaningful human oversight, international humanitarian law could codify a specific obligation for human verification of the civil status of every fixed infrastructure included on a list of targets, with an expiry date beyond which the information becomes invalid and must be renewed using direct sources. This is a minimum standard. Applied to Minab, it would have required someone to ask whether that building was still what the records stated it to be.
Technology can determine how a weapon reaches its target, whilst it is up to humanity to decide whether that target should be struck. The question underlying all this is who, in practical terms, exercises that humanity, with what tools of knowledge and within which institutions. In Minab, someone should have known that there was a school there. The fact that nobody knew, or that nobody asked, is the point from which we must start again.
By Kit KLARENBERG
Contact us: info@strategic-culture.su
Private messages show show Hamish De Bretton-Gordon, a top British pundit on the Salisbury poisoning incident, was a close associate of the MI6 officer who handled the Russian intelligence defector at the center of the scandal — and reaped considerable benefits from his proximity to the case.
This September, a former British Army officer named Hamish De Bretton-Gordon filed a column with the UK Telegraph complaining that “Russian bots” had “tried to make out that I’m part of a security service conspiracy.” Having shot to prominence as a media commentator on chemical and biological weapons after the alleged poisoning of turncoat Russian spy Sergei Skripal, De Bretton-Gordon dismissed accusations that he collaborated with UK intelligence as the product of the “Kremlin’s malign tactics.”
Despite his fervent denials, leaked emails reviewed by The Grayzone indicate that De Bretton-Gordon has spent years privately coordinating his public statements with senior British military and intelligence figures. The correspondences show him training operatives in Turkey and instructing Israeli military figures on Syria’s still-elusive stockpile of chemical weapons, “and a bit of Novichok,” referring to the Russian toxin supposedly used to poison Skripal.
They also suggest he stood to profit personally from exaggerating the threat posed by Russia. In the emails, De Bretton-Gordon described his day job as “flogging gas masks” to British military and law enforcement. He seemingly grew his firm’s business with outlandish warnings of imminent Novichok hand grenade attacks by Russian agents. As his prominence grew, the BBC hired De Bretton-Gordon to serve as an advisor on a drama about the Skripal poisoning.
His principal contact listed in the emails is Pablo Miller, an MI6 officer and veteran of the Royal Tank Regiment. Miller is widely believed to have recruited and handled Skripal before the Russian GRU defector was allegedly poisoned in Salisbury, England in March 2018 under questionable circumstances.
After the Salisbury incident, the British government issued a D-notice to prevent mainstream reporters from investigating Miller’s relationship with Skripal. Despite being an obvious subject of interest, Miller went completely unmentioned during the rolling coverage that dominated British media for months as a result.
The emails show Miller was nonetheless following that coverage closely. In June 2018, he wrote to De Bretton-Gordon thanking him for his “sterling media work” during what he called the “dark days of March and April.” During that period De Bretton-Gordon appeared in British newspapers and television reports on the poisoning on a near-daily basis, consistently attributing blame to the Kremlin without evidence and emphasizing the chemical weapons threat he claimed Russia posed.
The next month, De Bretton-Gordon sent his MI6 pal a lengthy Telegraph article warning of potential all-out chemical and biological wars with “rogue” states. De Bretton-Gordon’s commentary featured extensively in the piece, which framed Britain as a sitting duck in the event that Moscow were to target the country with missiles equipped with nerve agents or other toxic weapons.
When not hyping up De Bretton-Gordon’s apocalyptic warnings of potential Russian chemical weapons attacks, the article functioned as a long-form infomercial for his then-employer, Avon Protection Systems. The firm, which manufactures protective equipment for military, law enforcement and fire personnel, stood to profit handsomely from inflated perceptions of what governments call Chemical, Biological, Radiological, and Nuclear (CBRN) threats. De Bretton-Gordon’s email signature at the time identified him as the company’s Managing Director for CBRN.
De Bretton-Gordon’s big break
The Telegraph noted the company’s business was “booming,” with orders “flooding in” from government clients the world over, supposedly reflecting how “a new age of poison weapons is upon us.” Later in the piece, an Avon executive boasted that the Salisbury incident and other recent alleged CBRN usage had “been good for us.” Britain’s elite SAS were among the cited customers for the “sinister looking” gas masks sold by Avon Protection Systems, which were said to be “worthy of Darth Vader.”
At no point did the Telegraph article mention De Bretton-Gordon’s employment by Avon.
But thanks to his newfound media prominence, De Bretton-Gordon wouldn’t have to stay with the company much longer. In January 2020, he emailed Pablo Miller bragging that he had “been offered a fellowship” at Cambridge University. De Bretton-Gordon wrote that he’d decided to accept the role because, in his words, “it beats flogging gas masks!”
Four months later, De Bretton-Gordon founded a dedicated personal “CBRN and bio security” consulting firm, in an apparent attempt to derive profit from his talent for stoking public fears of looming chemical and biological catastrophe.
De Bretton-Gordon helps MI6 officer avoid media scrutiny
The leaked emails show De Bretton-Gordon quickly emerged as a preferred adviser for the BBC, and repeatedly served as a go-between for Miller and the British broadcaster, whose journalists had apparently become aware that Miller was an MI6 officer.
Correspondence from November 2018 shows De Bretton-Gordon was also rewarded for his relentless media appearances with a position as a “senior CBRN mentor” in Britain’s Ministry of Defence. Though the work was unpaid, he received an MOD ID card and Ministry railcard, saving him substantial sums on train travel. De Bretton-Gordon wrote Miller to inform him of his new title, noting that he was currently advising the BBC on what he derided as “CBRN ‘pisch’” for a forthcoming “drama on Salisbury.”
“I’ve been asked by the writer if you would be interested in contributing,” he mentioned to Miller, adding that he’d told the British broadcasters that the spy was unlikely to agree. The MI6 veteran responded, “you correctly guessed my answer to the BBC question… it would not be appropriate for me to comment in public on the issue.”
Miller lamented that the British state broadcaster’s cognizance of his clandestine involvement with Skripal indicated his “true professional provenance” had “long ceased to be much of a secret.” But he wished De Bretton-Gordon good luck, telling his colleague that he was sure the drama would succeed given his “expert guidance.”
The next January, De Bretton-Gordon explicitly described Skripal as Miller’s friend when forwarding him a Forces News article on the removal of the roof of the Russian defector’s former home, excitedly informing him, “I see your mate is in the news again!”
De Bretton-Gordon was quoted liberally in the report, explaining it was necessary to remove the roof as it had allegedly accumulated “enough Novichok to kill thousands of people.” The chemical warrior claimed British authorities were now “much better prepared to deal” with future CBRN attacks in the aftermath of the incident.
Weeks later, De Bretton-Gordon emailed Miller to warn him Salisbury “will be awash with journos” on the March 4 anniversary of Skripal’s poisoning. In response, Miller thanked his friend for the “heads-up,” and wrote that he’d likely be leaving the country to ensure he was far away from the city while it filled up with reporters.
The correspondence also reveals how close the two men had become, and how their relationship had begun to blur into commercial territory. In October 2019, Miller asked De Bretton-Gordon to help publicize a Kickstarter for his daughter’s theatre production — a personal favor suggesting a friendship rather than a purely professional engagement. But Miller refused to let media contacts anywhere near it, explaining that following the Salisbury incident his daughters were forced to close their social media accounts after being relentlessly hounded by the press.
During the same exchange, De Bretton-Gordon reminded Miller he’d been contracted for an upcoming BBC drama on Salisbury, and once again offered him a role. Miller declined, saying he was firmly opposed to playing “any part for any BBC show. Ever.” De Bretton-Gordon defended his participation in the project by revealing that the MOD had specifically requested he “help on the Salisbury piece… to make sure the military bit looks right.”
The production in question was a three-part BBC One series that aired in 2020 called “The Salisbury Poisonings.” The project’s IMDB page lists De Bretton-Gordon as a “military advisor.”
De Bretton-Gordon wants critics “prosecuted for treason”
As The Grayzone previously revealed, De Bretton-Gordon played a deeply malign, multifaceted role throughout the Syrian dirty war. On top of endless media appearances as a bogus “expert” on CBRN matters, he was part of a covert MI6 effort to smuggle soil samples out of opposition-controlled areas in an effort to prove the Syrian government used chemical weapons. A report on this initiative in the Mirror tabloid made clear the purpose of British intelligence was prodding the US into military intervention against Assad.
It seems plausible whatever samples De Bretton-Gordon gathered in Syria were concocted, perhaps by his own hand. He provided evidence to a number of official investigations into purported chemical attacks during the conflict. In at least one instance – an OPCW/UN Joint Investigative Mechanism (JIM) probe into an alleged April 2014 toxic gas strike in Talmenes – videos submitted by an opaque NGO De Bretton-Gordon founded called CBRN Taskforce were found to show clear signs of falsification.
De Bretton-Gordon’s communications with his spy buddy contain references to him visiting Turkey, “to do a bit of training.” The country was the central hub for MI6 and CIA operations during the dirty war on Syria. It is an open question whether his training sessions covered staging chemical weapons attacks, and if other states in the region were taught in this dark art.
The self-styled CBRN expert also bragged about sharing his purported expertise with the Israeli military. In an email sent in December of 2018 De Bretton-Gordon boasted to Miller that he’d just returned “from Israel,” where he said he’d been “briefing the IDF (and no doubt others!) on Syrian [chemical weapons] and a bit of Novichok.”
He was clearly impressed by the Israelis, crowing, “Those boys don’t mess about!”
De Bretton-Gordon’s work overseas was apparently taking a toll on his day job, however. In an email that month to Miller, he explained how he was “always on the edge of sacking with all my ‘extra’ activities.” These machinations also put him in conflict with Peter Ford, Britain’s ambassador to Syria from 2003 to 2006, who’d become an outspoken critic of the dirty war against Damascus.
After De Bretton-Gordon authored an op-ed for The Telegraph in May 2019 demanding the West intervene to prevent Syrian government forces from recapturing Syria’s Idlib, Ford issued a response nailing his failure to acknowledge that the province was at the time under the control of the MI6-supported Al Qaeda affiliate, Hayʼat Tahrir al-Sham.
In response, De Bretton-Gordon emailed Miller a note deriding Ford as a “complete See You Next Tuesday” who parroted “Assad and Putin propaganda.” He called for Ford – and the members of a now-defunct collective of antiwar academics known as the Working Group on Syria, Propaganda and Media – to be “prosecuted for treason.” De Bretton-Gordon seemed particularly piqued that the academic collective had poked holes in the narrative he’d spun about Syrian chemical attacks.
De Bretton-Gordon cheers as his enemies “exposed”
In August that year, De Bretton-Gordon contacted Miller to celebrate The Times publishing a hit piece on David Miller, a member of the Working Group on Syria, Propaganda and Media.
“Glad to see that some of Assad’s and Putin’s ‘useful idiots’ who try and give me a hard time are being exposed,” De Bretton-Gordon cheered, before suggesting “must get more of them out in the open.”
This prompted a wider discussion about the so-called “stooges” who publicly challenged established, MI6-approved media narratives on Syria and Russia.
For his part, Pablo Miller expressed disgust over a recent “fawning evaluation” of Putin’s 20 years in power by dissident journalist Mary Dejevsky. Miller derided Dejevsky as one of many media figures who foolishly believed “Putin was the answer to Russia’s woes,” and suggested without evidence that she was likely “taking Moscow gold.” Thankfully, he added, events “Crimea, Syria, and Salisbury” had managed to “wake the twits up.”
The slanted manner in which these events were framed in the media did indeed coincide with a perceptibly negative shift in public sentiment towards Russia throughout the West, helping make a world-threatening standoff with Moscow a fait accompli. At every step, De Bretton-Gordon and other intelligence-aligned commentators acted as military and intelligence propagandists, ensuring the media and popular culture was constantly deluged with paranoia over a multi-spectrum Russian “threat.” Those who question the narrative are denigrated as “useful idiots,” and much worse.
This August 6, De Bretton-Gordon based an entire op-ed for the Telegraph on a conversation he’d recently had with an unnamed “Western intelligence officer” of his acquaintance. Could that anonymous spy have been Pablo Miller?
The self-styled chemical warfare expert was clearly not keen to discuss his relationship with the MI6 man. De Bretton-Gordon immediately hung up the phone when I reached him and identified myself as a journalist.
Original article: thegrayzone.com
Portugal must capitalise on its energy resources and digital infrastructure to attract investment in artificial intelligence (AI) and help Portuguese businesses expand internationally, according to Gonçalo Regalado, CEO of the
The post Portugal must catch up in AI race, warns development bank chief appeared first on Portugal Resident.
When people think about what ChatGPT or other large language models are doing behind the scenes, it’s common to reach for the language of computing. You might assume there is a database, or a logical series of steps being executed, in order to provide a response to your query. Picturing AI responses as akin to […]
The post Mind games: AI behaves more like a brain than a database appeared first on Asia Times.
We are fortunate to live at an incredible moment in history. As more and more AI applications have become available, people have turned to them to seek advice, build businesses, express ideas, learn, and improve their health and quality of life. Nowhere is this more clear than in Asia. AI is already turbocharging small businesses […]
The post Choices that will define Asia’s AI future are coming into focus appeared first on Asia Times.
Russian forces appear to use AI to plan air attacks and recognize targets amid raids on Ukrainian cities. This may include recent air attacks on Odesa that have come in waves with short gaps, Deputy Air Force Commander for Air Defense Colonel Yurii Cherevashenko says, Ukrinform reports.
Russia has already equipped its jet-powered Geran-4 drones with machine vision, allowing them to recognize and track targets on their own in the final seconds of flight. Recently, a single jet drone hit a locomotive in Odesa Oblast in late September and killed two railway workers.
The scale of Russia's AI use is hard to measure, Cherevashenko continues, but Ukrainian specialists already see its elements in how attacks are organized. The long, wave-by-wave strikes on Odesa would be far harder to plan and coordinate without it, he believes.
"Russian forces use AI to lay out attack routes and to recognize targets," Cherevashenko says, framing it as his own assessment rather than a confirmed fact.
Odesa Oblast is home to the biggest port in Ukraine, "Southern". Protecting it is a challenging task as guarding the airspace over water is tougher than over land, Cherevashenko says. Some air defense assets do not perform as well at sea, and in some cases, crews can intercept a target only as it approaches.
Weather adds to the problem, with storms and fog, while the open sea gives Russia more freedom to choose drone routes and launch points. To protect ports and the coast, Ukraine uses drone motherships, platforms that carry other drones aloft to extend their flight time and keep watch over set areas.
Ukraine is also building its own AI for air defense. Automated turrets using machine vision now help counter Russian drones. According to Army Media, the Air Force says such turrets have scored their first shoot-downs, including of Geran-5 jet drones, with one soldier saying 50 rounds were enough to bring down a Geran.
Ukraine's Defense Ministry is going further. In late September, it unveiled an AI system called SPECTR, built by its A1 Artificial Intelligence Center and already in testing with several combat units, to analyze battlefield data, predict Russian moves, and suggest options to commanders.
Researchers in Porto are testing a system that uses drones and artificial intelligence to spot early signs of deterioration in railway bridges, helping operators plan repairs before problems become critical.
The post Porto researchers use AI and drones to spot railway bridge defects appeared first on Portugal Resident.
Giorgia Meloni has applied to register her voice with the European Union Intellectual Property Office as governments and public figures look for new ways to counter increasingly convincing AI-generated deepfakes.
The Italian prime minister submitted a short audio recording in which she says “Io sono Giorgia Meloni” — “I am Giorgia Meloni” — to the EUIPO on Monday.
The four-second recording forms part of an application for a sound trademark intended to strengthen legal protection against unauthorised use of her voice.
The application filed with the EU intellectual-property system is currently listed as under examination. It is linked to categories including downloadable multimedia material, cultural activities and the organisation of political events.
Registration would not give Meloni ownership over every recording or imitation of her voice, nor would it automatically prevent AI systems from reproducing it.
It could, however, provide another legal basis for challenging certain commercial or misleading uses where the protected sound mark is employed without authorisation.
The distinction between a trademark and copyright is important in this case. Meloni has applied to protect a sound identifier through EU trademark law rather than to establish general copyright ownership over the natural characteristics of her voice.
Her application comes as voice-cloning technology becomes increasingly accessible.
Modern generative-AI systems can create convincing synthetic speech from relatively small samples of recorded audio. The technology is used legitimately for dubbing, accessibility and media production, but it can also be used for fraud, impersonation and political disinformation.
Meloni herself has previously been targeted by manipulated images and other deepfake material. She has also taken legal action in Italy over sexually explicit deepfake images created using her likeness.
The problem has acquired additional importance for politicians because synthetic recordings can be produced and circulated rapidly during election campaigns.
A false recording no longer requires sophisticated audio-editing equipment or extensive access to the target. Public figures provide large quantities of clean audio through speeches, television appearances, podcasts and social-media video — precisely the material needed to train voice-cloning systems.
The EU has already begun imposing specific transparency requirements on synthetic content.
EU Today reported in August that new provisions under the EU AI Act require certain AI-generated or manipulated audio, images and video to be marked in machine-readable form and impose disclosure obligations for deepfake material.
Those rules address transparency rather than ownership.
Trademark law potentially offers a different route by allowing an individual or organisation to claim rights over a specific sound used as a commercial or identifying mark.
Sound trademarks are not new. Companies have long registered jingles, tones and other distinctive audio signatures. What is unusual in Meloni’s application is the attempt by a serving head of government to use the system partly as protection against AI impersonation.
The phrase chosen for the application is itself closely associated with Meloni’s political identity.
“Io sono Giorgia” became widely recognised following a 2019 speech in which she described herself as a woman, mother, Italian and Christian. The phrase subsequently became the basis of online remixes before being adopted as the title of her autobiography.
Other performers have also sought stronger protection for their voices as generative AI improves.
Italian singer Giusy Ferreri has filed a similar application, while actors and musicians internationally have pressed governments and technology companies to prevent unauthorised voice cloning.
Meloni’s filing is unlikely on its own to settle the broader legal question of whether an individual can effectively control AI reproductions of his or her voice.
It does, however, illustrate how existing intellectual-property law is being tested against technology that can reproduce personal characteristics that were previously difficult to copy convincingly.
The EUIPO must now determine whether Meloni’s application satisfies the requirements for registration. Until that process is completed, the filing remains an application rather than an approved EU trademark.
Login