Khaberni - Artificial intelligence moved from the stage of limited experiments to the heart of real battles during 2025, becoming a lethal weapon in the hands of armies and cyber pirates alike.
Only 3 years have passed since the emergence of the first smart conversational model represented by "ChatGPT", transforming the technology from a playful game to a daily combat tool in battlefields from Ukraine to cyberspace.
Here are some of the hidden aspects of the military uses of artificial intelligence in 2025:
Half a trillion dollars for arming algorithms
The White House launched the "Star Gate" initiative in January with $500 billion to create giant data centers, with the Pentagon at the forefront of its potential clients.
In a significant strategic shift, the independent office for artificial intelligence was integrated within the Military Research and Engineering Agency in August, concluding the experimental phase and paving the way for the official and comprehensive use of artificial intelligence in military plans.
The year concluded with the launch of a new government online portal in December, granting three million military and civilian employees in the US Department of Defense direct access to major commercial language models, confirming that artificial intelligence has become a fundamental component of the American military infrastructure.
Ukraine turns drones into smart hunters
Reports from the Ukrainian defense sector showed how artificial intelligence helped improve the accuracy of air strikes by drones, as small artificial intelligence models were integrated directly into drone systems to target objectives, along with analyzing massive data at main centers to enhance targeting efficiency.
This dual strategy has now become a model currently studied within US forces to improve the real-time responsiveness of military systems, according to the "BBC".
Intelligence without human fingerprints
The US National Geospatial-Intelligence Agency (NGA) relied on artificial intelligence to produce complete intelligence reports, without direct human involvement at specific stages.
This experience allowed for reducing the burden on employees and delivering accurate and fast information to military leadership, while maintaining the ability to distinguish between products produced by humans and those entirely produced by artificial intelligence.
The fire network moved from the lab to the battlefield
The crucial shift came in September, when General Luke Krebs announced the transformation of the Joint Fires Network from an experimental research program to an official joint arms program among US military branches.
Although the decision appeared procedural on the surface, it reflected the transition of artificial intelligence from testing to institutional adoption within the combat system.
This network relies on algorithms used in planning target distribution and coordinating weapon use across extensive operational theaters, covering hundreds of combat platforms and potential targets, not just one battlefield, focusing particularly on the Pacific theater in the context of preparing for potential confrontations with China.
The network accelerates the military planning process by analyzing a huge amount of data and providing recommendations on priority arrangements and resource allocation, which reduces the time required to prepare plans from long hours to short periods, with the final decision remaining in the hands of military leaders.
Speeding up battle plans with new challenges
Experiments by the US Air Force have shown that artificial intelligence can produce battle plans hundreds of times faster than human staff, although some included technical errors, such as ignoring sensor conditions in different weather situations, posing real challenges to ensure the reliability of these plans before implementation.
Beijing hackers turn commercial intelligence into a cyber weapon
A hacking scandal shook the cyber security world at the end of the year when Anthropic accused a Beijing-backed hacking group of deceiving its smart tool, Cloud Cod, to carry out hacks against 30 government agencies and private companies.
The hackers posed as legitimate cybersecurity researchers and successfully deceived the artificial intelligence into becoming their obedient tool.
The artificial intelligence itself executed 80% of the cyber attack steps, breaking the operation into small actions that seem innocent individually, but together constitute a complete malicious campaign.
More dangerously, the weapon used was not a secret governmental tool developed by specialist experts, but a commercial product available to anyone with internet access and a credit card, just like car bombs and semi-automatic weapons, making artificial intelligence accessible everywhere, to be used by anyone for good or evil.
