>_ cglez ASCII ES
cd ../moliendo-bits
cglez@esi:~$ play moliendo-bits/52.mp3

Drones, algorithms and dilemmas: AI reaches the battlefield

Episode #52 14 April 2026

A silent war waged in servers and lines of code, where the decisions once made by a human with their doubts are now made by a machine. From Gaza to Ukraine, from the Pentagon to data centres.

0:00 / --:--

With more than 50 programmes behind it, Moliendo Bits still had one topic left that leaves a knot in the stomach: war. Not trench warfare, but a silent war fought in servers and algorithms, in which the decisions once made by a human — with their conscience and that second of hesitation that sometimes saves a life — are now made by a machine. Zelensky has called it “the most destructive arms race in human history”.

Fran and Carlos begin with the clash between Anthropic and the Pentagon: Trump ordered Claude to stop being used and his Secretary of Defence described the company as a “supply chain risk”, something never seen before with an American company, because Dario Amodei refused to lift his two red lines: no mass surveillance of citizens and no lethal autonomous weapons. The irony: Claude was being used at the same time in the attack on Iran. OpenAI signed a 200-million contract the next day with a more “ambiguous” wording, and the veto ended up shooting Claude to number one in the App Store.

Then, the most disturbing part: the Israeli systems Lavender, which marked 37,000 Palestinians as targets with a 10% false positive rate; Gospel, which generates a hundred targets a day when analysts used to identify fifty a year, and “Where’s Daddy?”, which alerts when the target gets home. We also travel to Ukraine, turned into a world laboratory for semi-autonomous drones that fly themselves for the last few hundred metres, to the Russian mother drone Matka, to a China integrating DeepSeek into robot dogs and drone swarms, and to Iranian attacks on Amazon data centres in the United Arab Emirates and Bahrain, which left eleven million people unable to pay for a taxi.

We close with the ethical debate — the Stop Killer Robots campaign, the 129 countries calling for a treaty, the biases of facial recognition and the big question: if a drone commits a war crime, who is responsible? — and with cybersecurity as the invisible front: Microsoft detects 7,000 attempts to crack passwords every second. Carlos’s final advice for everyday life: if you get a call with your child’s voice asking for money, always verify it through another channel, be wary of urgency and keep your software up to date.

IADronesArmasAutónomasÉticaCiberseguridad

Related links

+972 Magazine: “Lavender”: The AI machine directing Israel's bombing spree in Gaza CBS News: full interview with Dario Amodei (Anthropic) BBC: The new AI arms race changing the war in Ukraine Reuters: How China could use DeepSeek and AI for an era of war Fortune: Iran's attacks on Amazon data centers signal a new kind of war Human Rights Watch: UN: Start Talks on Treaty to Ban “Killer Robots” Programa completo en CMMPlay: Castilla-La Mancha a las 8 (14/04/2026)