Key Insights
Higher education is uniquely vulnerable to lateral attacks.
Across the 159 million attacks analyzed in Abnormal AI's 2026 Attack Landscape Report, higher education stood out for the high share of attacks arriving from inside the institution. Nearly one in eight phishing emails reaching students comes from a compromised internal account, more than seven times the next-closest industry. The business email compromise picture is even starker: a third of all BEC targeting higher education comes from a real internal account rather than a spoofed one.
That pattern explains much of what makes university email hard to defend. A message sent from a .edu address passes authentication and conveys an immediate sense of legitimacy. There's no technical red flag, and no social one either.
Why Higher Ed Is So Exposed
Much of higher education’s vulnerability to identity-based attacks comes down to structure. Higher education campuses function like small cities, holding everything from intellectual property and sensitive research data to detailed student records. Massive amounts of money flow in through federal grants, donations, commercial operations, and athletic programs. The attack surface is equally broad, encompassing students, faculty, staff, vendors, contractors, and alumni.
This complex environment operates inside a culture of academic freedom. Universities prioritize open communication and the free exchange of ideas. Meanwhile, faculty directories, financial budgets, and grant awards are often published for anyone to read. Bad actors scrape those details within minutes to build highly convincing attacks.
These conditions mean that attacks reaching campus inboxes today look more like normal communication than anything malicious. And that makes them hard to stop with legacy, signature-based tools.
Ghost Students: Higher Ed’s Billion-Dollar Problem
Ghost students are aliases or stolen identities that fraudsters (and the bots they run) use to get accepted into a college. The goal is to access financial aid under a fabricated identity.
The fraudster completes an online application and, once accepted, enrolls in classes. From there, the fake identity gains the same access as a legitimate student, allowing the threat actor to abuse cloud storage or use the .edu email address to run other scams.
The major goal of ghost student fraud is to access federal grants funded by taxpayers. The Education Department says it prevented $1 billion in student aid fraud in 2025 alone. Earlier this year, FinCEN, a bureau of the U.S. Treasury, flagged ghost students as a growing threat to the $120 billion of aid disbursed each year.
That scale is only possible because applicant tracking software, built to move applicants through a pipeline, is not designed to check whether those applicants are real. North Korean state operatives have run this playbook against employers for years, using stolen identities to land real US jobs and route salaries and network access to the regime. Higher education is now seeing the same technique aimed at admissions.
Student-Targeted Phishing Scams
In recent years, phishing has shifted from a numbers game built on high-volume, easily recognizable scams to highly automated, hyper-personalized attacks powered by generative AI and real-time execution.
The student context gives phishing attackers plenty to exploit, from academic goals and career ambitions to financial pressures, housing concerns, and looming deadlines. AI tools can turn those details into natural, highly convincing messages, such as a prestigious internship offer requiring a modest application fee.
Attackers can also impersonate the finance office, housing, and enrollment, timing their messages around tuition deadlines, aid disbursements, and registration windows when urgency is already high.
These attacks pass authentication because, quite literally, they look authentic. Higher education's entire operating model depends on unfamiliar people emailing each other credibly, and each of those roles gives attackers cover to hide suspicious requests inside everyday traffic. The only distinguishing feature of a phishing email is that the sender is not behaving normally.
The Growing Risk of Shadow AI
A decade ago, IT teams battled shadow IT, the personal devices connected to university networks without approval. Today, they are tackling the even bigger challenge of shadow AI: the use of ungoverned AI tools.
Shadow AI has quickly become a runaway governance issue in higher education. In early 2026, more than half (56%) of the professionals surveyed by EDUCAUSE, the higher education technology association, said they had used AI tools for work that weren't provided by their institutions.
From a cybersecurity standpoint, the biggest concern with shadow AI is data leakage. Users routinely feed confidential research, student data, and IP into ungoverned platforms without reading the terms of service. Once data is out, you cannot pull it back. And there’s always the risk of an autonomous agentic AI creating a highly dangerous backdoor into a college network.
Some campuses are making approved AI usage easier by negotiating enterprise agreements with providers like OpenAI or Anthropic. But tighter controls rarely stop shadow AI, and exposure compounds as trusted applications add AI features of their own, sometimes changing their terms of service without anyone noticing.
Secure Email Gateways Weren't Built for These Threats
No institution achieves perfect security. The realistic goal is to reduce the number of opportunities attackers have to succeed and make the institution a harder target to compromise. But that’s not what the old model of defense was designed for.
Secure email gateways were built to detect obvious signs of an attack, such as malicious links or attachments. Because these tools focus on external inbound mail, they have no visibility into lateral movement from compromised internal accounts.
Today, higher education's dominant attack is identity-driven: an email sent from a real, trusted account, asking for something that sounds completely normal. There’s nothing for a signature-based check to catch.
Behavioral AI: A Different Starting Point
The solution is not to establish whether a message is technically “clean,” but to ask whether the account behind it is behaving the way it always does.
Abnormal’s Behavioral AI profiles every person, partner, and app connected to an institution by recording normal login times, sign-in locations, and device usage. It studies normal conversation cadence, tone, and sentiment to understand how people interact with each other.
The platform analyzes roughly 50,000 of these signals per email. When a message deviates from the expected “known good” behavior, the platform flags it instantly, with automated actions to remove suspicious messages from inboxes, remediate compromised accounts, or alert security teams.
Solving the Ghost Student Problem
Abnormal's Infiltration Prevention module applies the same behavioral engine to the admissions process. Originally built to catch nation-state actors faking, and sometimes deepfaking, their way into corporate jobs, it sources data from systems like Canvas and Instructure to profile higher education applicants and flag fabricated identities before they can access grant money.
Shutting Down Shadow AI
Abnormal extends that visibility beyond email with AI Governance, a module that helps security teams identify and control unapproved AI use across the institution. When a user signs up for a new AI tool, Abnormal can detect the confirmation email and alert IT before the signup is complete. Security teams can then block the signup, warn the user that the tool isn't approved, or allow it to proceed.
A lightweight browser plugin adds visibility into how AI tools are being used, including which services employees access and what data they upload. If an approved application later adds AI features and changes its terms of service or data processing agreement, the module catches it automatically.
All of this sits in the same console as email security and security awareness training. Most customers get a full view of the ungoverned AI running in their environment within 48 hours.
Next Steps for Higher Ed Security Leaders
It may feel disruptive to change an established security stack, but tools built for the old ways can't catch today's attacks. Institutions relying on legacy tools fall further behind every day.
For higher education security teams, a good next step is to run a proof-of-value assessment. Abnormal connects to an email environment in read-only mode within minutes, without disrupting existing tools or workflows.
The assessment gives security teams a clearer view of their current exposure and helps answer questions such as:
How much malicious email is getting through?
How many phishing reports are false alarms, and how much staff time do they consume?
Is security awareness training changing behavior or simply satisfying a compliance requirement?
How many fraudulent applicants, or “ghost students,” could already be in the admissions pipeline?
How much shadow AI exists across the environment?
Addressing those risks doesn't require more headcount. Abnormal automatically triages user-reported phishing emails and delivers security awareness training, reducing the amount of manual work for lean security teams. It also detects and remediates novel attacks automatically, helping teams respond faster and stop threats before they reach students or staff.
Higher education’s threat landscape is changing quickly, and these risks are only part of the picture. See what Abnormal’s latest research reveals about the attacks targeting colleges and universities—and how they compare with threats across other industries.
Explore the Higher Education Threat Landscape