Italy’s cyber command braces for AI-powered attacks as Milano Cortina Winter Olympics near

January 28, 2026
Italy’s cyber command braces for AI-powered attacks as Milano Cortina Winter Olympics near

ROME, Jan 28, 2026, 16:01 CET

  • Italy’s National Cybersecurity Agency faces its first big challenge with the Milano Cortina Winter Games running from Feb. 6 to 22, marking its debut since launching in 2021
  • Officials anticipate threats from petty cybercriminals to state-linked groups, with AI tools probably speeding up the pace of attacks
  • ACN is set to connect its specialists directly to the Games’ central tech hub in Milan, working alongside Deloitte and the organisers

Italy’s National Cybersecurity Agency has spent the last year monitoring criminal activity on the dark web and is now gearing up for AI-driven cyber threats linked to next month’s Milano Cortina Winter Olympics. Around 20 of its roughly 100 operational specialists will dedicate themselves full time to Olympic intelligence from Rome, while 10 senior experts will be stationed at the Games’ Technology Operations Centre in Milan, working alongside Deloitte and the event organisers. (Reuters)

The Winter Games take place from Feb. 6 to Feb. 22 and, for the first time, will be spread across several Alpine regions. This expansion complicates logistics and security efforts. With more locations involved, there are additional connected systems, suppliers, and potential failure points that could create major disruptions.

Rear Admiral Gianluca Galasso, ACN’s head of cyber operations and crisis management, warned that Italy is gearing up for roughly three billion viewers plus 1.5 million ticketed attendees. “That kind of exposure draws criminal interest,” he said, calling the Olympics “a stage” where cyberattacks often link back to political or geopolitical motives. (The Independent)

Analysts in the agency’s operations rooms in Rome monitor both public websites and secretive criminal forums for signs of upcoming attacks—whether stolen passwords or talk of fresh malware. The dark web, a hidden and encrypted part of the internet, serves as a marketplace where criminals exchange data and hacking tools.

Galasso noted that artificial intelligence introduces a new angle, with attackers likely deploying AI agents—automated programs—to aid cyber operations. While the agency hasn’t identified a specific spike in risk so far, it anticipates the overall threat will increase as the Games approach.

Italy lent a hand to French authorities during the 2024 Paris Summer Olympics amid over 140 recorded cyber incidents, including 22 breaches that allowed attackers into information systems. None, however, disrupted the events. On the opening ceremony day, France’s TGV high-speed rail network suffered physical sabotage involving explosive devices. (SRN News)

For Milano Cortina, ACN is on alert for any disruptions that could steal the spotlight: streaming outages, blocked sites, or glitches preventing ticket sales. These issues might be short-lived, but they hit hard when the event is live and seen worldwide.

The agency is focusing on early warning—detecting hostile activity before an intrusion attempt even begins—and then immediately relaying that information to technical teams at the venues. Galasso described this approach as operating “before someone tries to enter a system.”

Early detection hits a wall, especially with venues spread out and tech partners scrambling to keep systems running. On top of that, AI tools might help criminals ramp up phishing and reconnaissance, making defenders wade through even more noise without losing track of genuine threats.

Galasso said ACN will continue sending alerts as threats change, emphasizing speed the moment a risk shows up. “The goal is always the same: anticipate threats and respond quickly and effectively,” he said.

Technology News

  • AI arms race in U.S. colleges as detectors clash with humanizers
    January 28, 2026, 10:06 AM EST. Across U.S. colleges, a surge in generative AI has turned classrooms into an arms race between AI detectors and tools called humanizers. Universities say hundreds of students were flagged for cheating, yet critics argue detectors are unreliable and biased against non-native English speakers. An expanding group of students say their work was falsely marked as AI-generated, with lawsuits following. In response, firms such as Turnitin and GPTZero have upgraded detection and rolled out browser- and writing-history tools to prove authorship. Some students rely on humanizers to evade detection; others insist they never used AI. Educators warn the cycle could outpace policy and reshape definitions of cheating and authorship.