Spy Turtles, AI Killers & Curfews 🐢

If you’ve ever suspected your fish was watching you a little too closely, China’s Ministry of State Security would like you to know: you might be right. This week, Beijing’s intelligence officials warned that foreign spy agencies have been attaching sensors to marine animals, including turtles and fish, deploying them in Chinese waters to collect data on ocean temperature, salinity, currents, and coastal defences. The animals reportedly transmit their findings via satellite. Beijing claims these “spy turtles” and “spy fish” are building underwater maps that identify weak points in China’s naval defences. No country has been named, and no evidence has been provided. But the image of a sea turtle wearing a sensor backpack and paddling dutifully past a naval base is almost certainly going to be the one for history books.

The United States now exports 10.5 million barrels of oil per day, making it the world’s largest oil exporter for the first time. Saudi Arabia is at 5.9 million barrels; Russia is at 7 million. The shift reflects a decade of shale expansion, but recent disruptions have accelerated the timing: sanctions on Russia, Ukrainian drone strikes on oil infrastructure, and the ripple effects of the US-Israel war with Iran have all squeezed Saudi and Russian output at precisely the moment American production hit new highs. The last time a single country so decisively shifted global energy markets was 1973, and the world looked very different then.

What It Is: Fully autonomous AI drones have killed human soldiers in combat with no human in the decision loop. A senior Ukrainian defence industry figure confirmed that in a test near Bakhmut and Chasiv Yar, 10 “Terminator” quadcopter drones were programmed to fly toward the front line and engage targets independently. They killed Russian soldiers and destroyed a truck. No human pressed a trigger or made the call to engage.

Why It Matters: When a machine decides to kill a person without any human authorising that action at the moment it happens, something fundamental shifts. The reaction from much of the West has been silence, or quiet approval, because Ukraine is doing it against Russia, and we know which side is which. But nationality is beside the point. The question of whether any AI system should be permitted to end a human life with no human sign-off applies regardless of the uniform the target is wearing. If we accept this when Ukraine does it against Russia, we have conceded the principle entirely. No matter who is being killed, no matter their colour, no matter their nationality, the decision to take a human life should involve a human.

The Big Picture: International conventions ban chemical weapons, cluster munitions, and landmines because humanity agreed, after witnessing their effects, that some tools of war are too indiscriminate and too dehumanising to permit. Autonomous lethal AI belongs in that category. The Campaign to Stop Killer Robots has been pushing for regulation for years; this test is the most concrete evidence yet that the debate is no longer theoretical, it has already happened. The hard question is accountability: when an AI kills the wrong person, who is held accountable? Right now, there is no answer. International agreement on this is not just overdue; the technology has already been deployed, while lawmakers are still pondering whether to regulate it.

What It Is: The UK government is moving to ban social media for under-16s and romantic AI chatbot apps for under-18s. The legislation, expected to formally launch by spring 2027, also restricts live-streaming and disappearing messages for under-16s, and late-night social media access for under-18s more broadly. Apple and Google have been given three months to implement nudity-detection filters on children’s devices or face legislation requiring them to. The scope goes well beyond Australia’s social media ban, which operated at the platform level.

Why It Matters: The government moved quickly once political will was in place, completing consultations and legislation in an unusually short timeframe. The reach is wider than anything seen elsewhere: no Instagram, no live streams, no disappearing messages, no AI chatbots designed to simulate romantic or sexual relationships, and no late-night scrolling. YouTube has pushed back, arguing that a blanket ban would deprive users of access to educational content and that its existing safeguards are effective. That debate is worth having, but the government’s position is that voluntary safeguards have had years to work and have not delivered. The device-level intervention is what sets this apart: asking Apple and Google to build nudity filters into their operating systems shifts enforcement from apps to the device infrastructure, which no other country has attempted at this scale.

The Big Picture: More than a dozen countries are watching this closely, with momentum building particularly across Europe. If the device-level approach works, it becomes the template. If teenagers route around it within weeks, as they have with every previous attempt, it will force either a harder legislative response or a quiet retreat. Political consensus around children’s online safety has hardened enough in several democracies to produce actual law, not just concern. Britain has now gone further than anyone else, covering social media, live streams, disappearing messages, and AI companions that previous frameworks didn’t address.

What It Is: France’s disinformation watchdog, Viginum, has identified BlackCore, an Israeli firm that describes itself as an “elite influence, cyber and technology” company, as the suspected operator behind a campaign targeting left-wing candidates in France’s 2026 municipal elections. This week, Viginum revealed that BlackCore is also suspected of meddling in elections in New York City and Scotland, and operating in Angola and Togo. Two separate investigations are now underway in France: one by Paris prosecutors, another by domestic intelligence. Who commissioned the campaigns remains unknown.

Why It Matters: One firm, apparently operating from Israel, ran influence operations across three Western democracies and two African states. The toolkit: fake accounts, amplification networks, and disinformation content routed through servers in Britain, Germany, Finland, and Lithuania. What BlackCore represents is a commercial service, not a state apparatus. AI and cheap computing have made disinformation something you can buy off a menu. What once required the apparatus of a government intelligence agency, targeting voters, seeding false narratives, amplifying at scale, can now be ordered from a company most people have never heard of.

The Big Picture: The Southport riots in 2024 and the Belfast violence that followed just recently were driven in large part by disinformation spreading faster than any correction could reach. Those were organic. What BlackCore represents is the professional, commercial version of the same threat. If one firm can allegedly interfere in elections in France, New York, and Scotland at the same time, the question for every democracy is the same: what stops the next one? As things stand, some European governments are moving at a snail’s pace in their efforts to regulate, tackle misinformation, and hold platform executives accountable. Meanwhile, the cost of running a disinformation campaign continues to fall. The more important story is where election interference is heading, and in my opinion, this is something which is the biggest threat to the national security of the West rather than any Russian or Chinese missile.

🇷🇺 Russia: Vladimir Putin has signed a law allowing the state to seize the property and bank accounts of Russians living abroad who are accused of offences against “Russia’s interests,” including posting criticism of the military online, before any trial takes place.

🇪🇺 EU: The European Commission has proposed barring any Russian who fought in Ukraine from entering the EU as part of its 21st sanctions package against Moscow.

🇳🇪 Niger: Niger’s military government has criminalised same-sex relations under a new penal code, with sentences of five to ten years in prison for same-sex relations and up to twenty years for same-sex marriage.

🇦🇫 Afghanistan: Taliban forces opened fire on more than 100 protesters in Herat after crowds gathered to demand the release of 30 women and girls detained for failing to meet hijab requirements. The protest is among the rarest instances of public dissent seen since the Taliban returned to power.

🇨🇭 Switzerland: St Gallen’s cantonal parliament voted 70 to 46 to ban female teachers from wearing religious symbols, including headscarves, in public schools.

🇨🇦 Quebec: Quebec has become the first jurisdiction in North America to ban the sale of energy drinks to under-16s, defining them as beverages with at least 150mg of caffeine per litre.

President Hakainde Hichilema has signed the Education (Amendment) Bill, 2026, making free schooling from early childhood through secondary school a legal entitlement for every child in Zambia. Since the policy was first introduced, 2.5 million children have returned to the classroom, and the government has recruited over 41,000 teachers in what is the largest education hiring drive in the country’s history. For a continent where financial barriers have historically pushed millions of children out of school before they were ready to leave, this is the kind of law that changes life trajectories.

“The real problem is not whether machines think but whether men do.”
— B.F. Skinner

Thank you so much for reading my curated news stories! If you'd like to support my work, a coffee goes a long way or share this article with a friend. Either way, thank you for being part of this. Your support means a lot x

Support My Work

Enjoying this newsletter? The biggest compliment you can give is to share it with your friends.

Share

Thanks for reading. You’re officially the smartest person in the room. Talk to you soon :)