For years, the cybersecurity industry has debated when the "AI Revolution" would truly hit the phishing landscape. According to the Hoxhunt Phishing Trends Report 2026, that moment arrived in December 2025. What began as a year described as "Cloudy with a chance of Skynet" transformed into a strategic "thunderhead" that has fundamentally shifted the baseline for email security.
Drawing on 50 million data points from more than four million users globally, the report is a mandatory read for CISOs and SOC teams. Here is a breakdown of the seismic shifts in the threat landscape and what they mean for organizations' resilience.
"Our research shows that AI-generated phishing went from a trickle to a flood almost overnight. The lesson for security leaders is clear: if attackers can use AI to scale social engineering, defenders must use AI to scale human cyber skills," said Mika Aalto, Co-Founder and CEO of Hoxhunt. "The biggest mistake companies can make in the AI era is believing technology alone will solve social engineering. Attackers are targeting human behavior. That means the defense must strengthen human behavior as well. The advantage will go to whoever understands that technology is a lever, not a replacement, for influencing human psychology."
Aalto added, "We've expected AI to reshape cybercrime for years, so the answer isn't panic, it's preparation. Right now, there's a wave of alarmist messaging around AI threats that almost resembles social engineering itself. Deepfakes are real, but they're still rare and highly targeted. If companies focus training on exotic attacks instead of the common social engineering tactics people face every day, they're not optimally managing human risk."
Until late 2025, AI-generated phishing was a niche tactic, representing only 1% to 4% of detected attacks. That changed during the holiday season. In December 2025, AI-generated campaigns surged by a factor of 14, accounting for 56% of all threats that successfully bypassed email filters.
Why this matters: These are not the "Nigerian Prince" emails of old. These attacks are high-volume, highly personalized, and architected to evade traditional detection.
43% of these AI attacks contain malicious links.
20% utilize open redirects to mask their true destination from filters.
11% deliver malicious attachments.
5% are "callback phishing" attempts, leading users toward malicious phone numbers.
One of the report's most tactical findings is the weaponization of the calendar. Phishing campaigns using .ics calendar invites are currently six times more dangerous than typical email phishing.
Because these invites automatically populate as meetings in a user's calendar, they create a persistent threat. Even if a user reports the original email as a threat, the calendar entry often remains behind like a "landmine," offering a second, long-lasting opportunity for a malicious click when the meeting reminder pops up later.
Attackers are increasingly moving away from broad "spray and pray" tactics to focus on high-value corporate accounts through recruitment scams.
The targets: Sales, marketing, and social media teams
The method: Impersonating major brands with fake job opportunities to harvest credentials
The goal: Hijacking corporate social media or advertising accounts, which can lead to significant brand damage and financial loss
Despite the darkening forecast, the report offers a clear path forward. The data prove that organizations can significantly reduce their risk by moving away from static, quarterly compliance training and toward adaptive security behavior change.
Reporting improvement: Companies using adaptive training saw a 6x improvement in threat reporting within just six months.
Click reduction: Malicious clicks were reduced by a staggering 86% to 87%.
The "human sensor": When employees are trained to recognize and report threats in real-time, they effectively become a decentralized security layer of "human sensors."
"We're seeing a widespread update to phishing. AI isn't creating completely new attacks yet. It is making traditional phishing campaigns more convincing, faster to produce and harder to detect. Agentic spear phishing isn't yet the new normal, but we're monitoring its development and creating countermeasures to prepare for its arrival," said Pyry Åvist, Hoxhunt CTO and co-founder. "Attackers are moving beyond email. Mobile phishing, callback attacks, and malicious calendar invites are examples of how social engineering is expanding into the everyday tools and workflows employees use outside of email. It's key that our technical and training protections are equally expansive."
Åvist continued, "The same AI systems that can craft highly-convincing phishing attacks are even more powerful when used for defense. If an AI agent can personalize attacks at scale, it can also personalize training at scale, helping every employee build resilience against those exact tactics."
As the report concludes: "Knowledge is power. By knowing our people, we have the power to become more secure."