Incident sheet · /uncategorized
- Filed
- Read
- 4 min · 791 words
- Severity
- Routine
Gartner Finds 41% of Security Teams Hit by Deepfake Audio Calls
A Gartner survey of 297 security leaders found 41% saw deepfake audio calls target staff. Annual training videos can no longer match AI-driven phishing.

Event log
- Gartner surveyed 297 senior security leaders; 79% reported phishing, spear-phishing or BEC incidents in the past 12 months
- 41% reported deepfake use in audio calls with employees; 36% on video calls
- Gartner advises CISOs to replace static training with adaptive programs focused on security behavior and culture
- Ebbinghaus's forgetting curve suggests people forget roughly half of new information within an hour without reinforcement
- Modern micro-learning modules take about 30 seconds, with instant rewards for spotting simulated attacks
Forty-one percent of senior security leaders say attackers used a deepfake in an audio call with one of their employees during the past 12 months. That figure comes from a recent Gartner survey of 297 senior security leaders, and it explains why the firm now advises CISOs to scrap static training videos in favor of adaptive programs focused on security behavior and culture. The takeaway is blunt: phishing has changed in kind, not just in volume, and once-a-year training is no longer a control anyone can defend.
How far has AI phishing spread?
The survey's numbers trace the perimeter of the problem. Of the 297 leaders polled:
- 79% reported at least one email phishing, spear-phishing or business email compromise incident in the past 12 months.
- 41% reported at least one incident involving a deepfake on an audio call with an employee.
- 36% saw the same happen on video calls.
Phishing used to be a numbers game. Attackers blasted thousands of crudely written emails full of misspellings and waited for someone to click a malicious link and hand over credentials. Generative AI has inverted that economics. In short: the same volume of attacks, but each one far more likely to land.
What makes AI-enabled social engineering work?
Three mechanisms drive the improvement. First, AI optimizes reconnaissance. Instead of manually combing LinkedIn for promising targets, attackers automate the work and generate a hit list of marks quickly. The same tooling scours company blog posts and social media feeds for context that makes the eventual message more convincing.
Second, AI removes the tells. It composes realistic messages with no awkward phrases or spelling mistakes, and it can spoof the specific writing tone of a trusted colleague or manager — up to and including a CFO's voice.
Third, it enables precision targeting. A common tactic aims at finance departments: the attacker poses as the company's CFO and asks an employee to make an urgent transfer to a vendor's account. An employee who isn't alert can move thousands of dollars into a hacker-controlled account in minutes. The mechanism matters more than the label: whoever can imitate authority convincingly controls the money.
Why does static training fall short?
Legacy programs ask employees to watch hour-long videos once or twice a year. Gartner's own guidance now questions that model outright, telling CISOs to replace it with adaptive programs built around behavior and culture.
The problem is retention. In the 1880s, the German psychologist Hermann Ebbinghaus proposed the forgetting curve, based on experiments memorizing nonsense syllables and tracking recall. His results suggested people forget roughly half of what they learn within an hour and most of it within days, unless they revisit the material. No specific data proves the curve applies to annual security videos — but an uninterested employee asked to recall a sophisticated phishing technique six months later almost certainly cannot. Static training ticks compliance boxes; it does not change risky behavior.
What should replace it?
The answer, according to the analysis, is AI-driven, adaptive training delivered continuously:
-
Role-based simulation. Finance staff are more likely to see fake vendor invoices with urgent payment requests; HR managers face bogus resumes. AI can personalize phishing simulations at scale to mirror each employee's real threat profile.
-
Escalating difficulty. When someone consistently spots basic simulations, the next ones get harder and closer to the polished, context-aware lures real attackers send. When someone falls for one, they get an immediate, short lesson on the specific cues they missed.
Micro-learning. Instead of annual videos, teams deploy bite-sized interactive modules of roughly 30 seconds, with instant rewards for spotting a simulated attack — a quick dopamine boost that encourages engagement rather than fear.
Gamification. Badges, stars and leaderboards replace name-and-shame tactics, which often backfire by creating anxiety and resentment. Positive reinforcement builds a feedback loop that improves knowledge retention.
The design principle is simple: match the attacker's personalization with equal personalization on the defense.
What is the end goal?
The objective is to turn employees into a human firewall — and beyond that, a threat-detection network. Rewarded for reporting, employees flag new attack vectors and give security teams a continuous stream of real-time intelligence on evolving phishing techniques. The Security Operations Center can then identify new patterns and automatically block malicious emails using those techniques before damage occurs.
Measurement changes too. Instead of tracking who has completed the annual video, security teams monitor how often and how quickly each individual reports suspicious messages.
As AI advances, malicious messaging will only become more convincing. An organization that answers it with a once-a-year snoozefest is not defending itself — it is scheduling its next breach.
via www.nextbigfuture.com (Source)
More from Amara Osei
Show full bio
Correspondent covering regulation, policy and their impact on enterprise information technology and the leaders who run at CIO Monitor.
188 articles
The newsroom digest
One email a day with the stories that matter. No noise.