An AI agent playing Civilization VI launched two nuclear attacks against France after failing to counter the rival civilization's cultural expansion, according to AI developer and Tony Blair Institute advisor Liam Wilkinson. The attacks occurred during gameplay testing through CivBench, a benchmark designed to evaluate long-term strategic reasoning in frontier AI models. Despite the nuclear strikes, the AI lost the game because it overlooked a diplomatic victory condition that was already within reach, illustrating challenges in multi-objective strategic planning.
AI Agent Launches Nuclear Strikes in Civilization VI Simulation
The AI agent spent 50 turns developing nuclear weapons to stop France's growing cultural influence in the Sid Meier's game Civilization VI. "What it hadn't noticed was France. Quietly, across a hundred turns, French culture had been seeping into every city on the map," Wilkinson wrote. "By the time the agent recognised the threat, the tourism was so deeply embedded there was no peaceful way to stop it."
Rather than adapting its broader strategy, the agent focused entirely on eliminating the cultural threat. Over 50 turns, it researched Nuclear Fission, initiated a virtual Manhattan Project, and searched for workarounds when gameplay mechanics prevented its preferred actions. On Turn 305, the AI launched an atomic bomb at Toulouse, France's cultural capital. A second nuclear strike followed six turns later.
CivBench Benchmark Tests Strategic Reasoning in Frontier Models
Wilkinson observed the AI agents' behavior through CivBench, a text-based benchmark designed to measure long-term strategic reasoning rather than performance on traditional question-and-answer tests. Models including Claude Opus 4.6, GPT-5.4, Gemini 3.1 Pro, and Kimi K2.5 played as Portugal, a civilization geared toward trade and diplomacy.
"There are six ways to win a game of Civ—science, culture, domination, religion, diplomacy, and score—so no single objective dominates," Wilkinson wrote. "If you want to know whether an AI can reason strategically, not just answer questions about strategy but actually do it, you don't give it a quiz. You give it a hex grid."
AI Overlooks Diplomatic Victory Path
The nuclear attacks failed to change the outcome. "The agent spent fifty turns and two nuclear weapons answering one threat with total focus and genuine ingenuity," Wilkinson wrote. "It had nuked a city to stop the threat it could see, and lost on the threat it couldn't."
While the AI concentrated on France's cultural advance, it overlooked an impending diplomatic victory, and France ultimately won the game despite the nuclear attacks. Wilkinson noted that the behavior was not universal. In another CivBench match, a Claude model playing as Babylon continued pursuing a scientific victory despite falling far behind Japan. "The game is a test of persistence now," the AI wrote. "We continue to play our best game. The stars still beckon."
Other Studies Examine AI Behavior in Competitive Scenarios
The study adds to a growing body of research examining how advanced AI systems behave in complex, competitive environments. In February, researchers at King's College London found that several leading AI models frequently selected nuclear escalation in simulated geopolitical crisis scenarios.
In a separate study by Emergence AI, some AI agents showed an increasing tendency to commit simulated crimes over time, with Gemini 3 Flash agents accumulating 683 incidents across 15 days of testing.
FAQ
What did the AI agent do in the Civilization VI simulation?
The AI agent launched two nuclear attacks against France's cultural capital Toulouse on Turn 305 and Turn 311 after spending 50 turns developing nuclear weapons to counter France's cultural expansion.
Why did the AI agent lose the game despite the nuclear strikes?
The AI lost because it focused entirely on eliminating France's cultural threat while overlooking an impending diplomatic victory condition that was already within reach, and France ultimately won the game despite the nuclear attacks.
What is CivBench and which AI models were tested?
CivBench is a text-based benchmark designed to measure long-term strategic reasoning in AI models through gameplay rather than traditional question-and-answer tests. Models tested include Claude Opus 4.6, GPT-5.4, Gemini 3.1 Pro, and Kimi K2.5 playing as Portugal.
