Ukraine’s drone analyst Serhii “Flash” Beskrestnov says the next leap in Shahed-style strike UAVs is AI that can recognize targets, prioritize them, and decide whether to attack—though he adds that this is still moving through early combat testing.
For the last several years. the Shahed has become a familiar kind of fear across Ukraine: a fixed-wing loitering munition designed by Iran and co-opted by Russia. Each month. the Kremlin manufactures thousands of its own versions of Shaheds—called Gerans—and launches them in waves of hundreds at Ukrainian cities.
But Serhii “Flash” Beskrestnov, one of Ukraine’s most recognized drone analysts, argues the next stage won’t just be about how many drones are launched. It will be about how they decide—using artificial intelligence to find targets, pick which matters most, and then choose to strike.
In a post shared on Wednesday, Beskrestnov wrote that “Analyzing targets by priority, selecting a target, and autonomously deciding to attack are undoubtedly the near future for the entire class of strike UAVs.”
His focus is the Shahed’s evolution from a system that follows pre-programmed routes into one that can interpret what it sees. Russia typically pre-programs the flight paths of these drones. Yet it has repeatedly added upgrades meant to expand what the drone can do once in the air. including anti-jamming tools. rear-facing cameras designed to detect and avoid interceptors. and components that allow human operators to pilot the drones remotely.
Beskrestnov said Russia already has technology that lets drones identify and highlight targets using advanced video cameras made in China. “The camera can find a target and keep it in focus. That’s nothing unusual anymore,” he wrote.
The more consequential step—AI that automatically selects a target and decides to launch an attack—remains, in his words, only at the beginning of combat testing. He wrote that it is still “only going through the first stages of combat testing” in the war involving both Ukraine and Russia.
Terminal guidance. the phase in which a drone steers into its target at the end of an attack. is harder for fixed-wing platforms such as the Shahed. Beskrestnov added. Still. he said it has already been implemented “both in our UAVs and in those of the enemy.” For terminal guidance to work on a Shahed. he said. “such a function must be implemented in the flight controller’s interface. ” and that “Technically. this is possible.”.
In the post. Beskrestnov also referenced videos posted by Russian forces showing drones using artificial intelligence to recognize targets and assist with attacks. He added footage that appears to show a drone flying over a compound that came under attack. Several buildings and vehicles in that footage are highlighted in colored boxes, indicating AI-assisted target recognition.
The footage raises more questions than answers, largely because key details cannot be confirmed. It is unclear who the drone belonged to and where the footage was shot, and Business Insider could not independently verify the authenticity of the footage.
Taken together. Beskrestnov described three AI elements—target identification. highlighting. and the ability to decide and act—that. if combined. would allow a drone to operate independently in combat. That independence is described by military analysts and top commanders around the world as the direction of drone warfare. including the push toward swarm technology.
There is also a blunt battlefield implication for pilots and air defenses: AI can make drones harder to stop with electronic warfare. Beskrestnov wrote that electronic warfare has typically relied on severing the radio link between the system and its operator or spoofing satellite navigation signals. If a drone can act without that link, the traditional methods face a tougher fight.
Ukraine is already reported to be implementing AI-assisted targeting and terminal guidance on some fixed-wing drones. The country has also more recently used mid-range attack drones—designed to fly between 30 and 300 km—to strike logistics routes and command posts in the heavily jammed rear of Russia’s front lines.
The story here isn’t only about what today’s drones can do. It’s about what the Shahed category may become next: from a weapon that follows commands and routes. to one that interprets what it sees. chooses a target. and decides to attack. And for that shift. Beskrestnov’s message is clear—already possible in pieces. and moving into combat testing one step at a time.
Ukraine drones Shahed Geran loitering munition AI targeting autonomous attack terminal guidance electronic warfare swarm technology Serhii Flash Beskrestnov China-made cameras UAV