The FZ1073 Cockpit Attack Ran Into Two AI Problems at Once: One Missing, One Already Running
Post-9/11 hardening stopped working the moment the threat moved to the left seat. A deepfake video impersonating BBC was circulating within hours. Neither was a surprise.
Explain Like I'm 5 — the whole thing in plain language.
A Flydubai flight from Dubai to Tel Aviv had to make an emergency landing on September 30, 2026. The co-pilot allegedly attacked the captain inside the cockpit. The captain was stabbed but managed to reach the cockpit door release. Passengers tackled the attacker. Two off-duty pilots who happened to be riding as passengers took over and landed the plane safely in Saudi Arabia. The captain is recovering.
This story touches AI in two ways, both bad, and different from each other.
The first problem: planes have a lot of security technology to stop someone from breaking into the cockpit from outside. That technology was built after September 11, 2001. It doesn't help when the threat is already sitting in the pilot's seat. Researchers have been writing papers about AI tools that could watch a pilot's behavior over time and flag unusual patterns early — using voice, phone signals, wearables. None of those tools are on any commercial aircraft today.
The second problem: within hours of the attack, a fake video using a BBC presenter's face — made with AI — was already spreading on social media, claiming things that weren't true. Separately, a wave of posts claimed the whole incident was staged. Both spread faster than real reporting could catch up.
Two AI problems on the same day. The defensive kind is still in academic papers. The harmful kind ships fast and needs no special skills to use.