Is KnowBe4 Training AI-Generated? KnowBe4 AI vs. AI-Native Platforms (2026)
Read summarized version with

Key takeaways
- KnowBe4 describes its core training library as "handcrafted, studio-grade content" from KnowBe4 Studios and uses AI for custom courses, phishing templates, translations, and training assignments, as of October 2026.
- KnowBe4 AIDA (AI Defense Agents) is included only with SAT Advanced among KnowBe4’s two current plans. SAT Foundation includes AI features such as AI-Selected Phishing Templates, AI-Recommended Optional Learning, and a Risk Score powered by the SmartRisk Engine.
- KnowBe4’s Content Creation Agent is limited to 25 generated courses per year, and the Synthesia license behind the Custom AI Video Builder comes at no added cost and covers 120 minutes of narrated video per year.
- AI-assisted security awareness training layers AI features such as recommendations, template generation, and agents on a human-produced library, while AI-native training uses AI to generate content, simulations, and personalization for each organization.
- Adaptive Security generates email, SMS, voice, and deepfake simulations from open-source intelligence (OSINT) on employees and builds custom training modules from prompts or policies with AI Content Studio.
On May 5, 2026, KnowBe4 announced an “AI-native” attack simulation and training product, and its AI-native campaign page carries the heading “Not Just AI-Assisted. AI-Native.” KnowBe4’s training library page describes the core content as “handcrafted, studio-grade content” from KnowBe4 Studios.
As of October 2026, KnowBe4’s training is partly AI-generated. The company describes its core library as studio-produced, and KnowBe4 AI generates custom courses, quizzes, phishing templates, and videos on top of it. AI-native security awareness training is a program in which AI generates the content, simulations, and personalization for each organization from the organization’s own data, people, and workflows.

Is KnowBe4 Training AI-Generated?
KnowBe4 training is partly AI-generated. On its training library page, KnowBe4 describes its core library as “handcrafted, studio-grade content” from KnowBe4 Studios. Its pricing page lists 200+ pieces of that content in SAT Foundation and 1,000+ in SAT Advanced, KnowBe4’s two security awareness training (SAT) plans. Beyond the library, KnowBe4’s AI agents and features generate custom content and decide which training and simulations each employee receives.
KnowBe4 applies AI to training in five ways, as of October 2026.
- Custom courses and quizzes: The Content Creation Agent builds text-based courses with a quiz from administrator prompts and internal policies. The Policy Quiz Agent generates quizzes based on an organization’s security and compliance policies.
- Phishing templates: The Template Generation and Callback Template Generation agents create phishing email templates and callback phishing templates on demand.
- Video: The Custom AI Video Builder produces AI-avatar training videos, and the Deepfake Training Content Agent produces a deepfake training video of a leader from the organization.
- Translation: Professional linguists and cybersecurity experts localize KnowBe4 content into 47 languages and dialects. KnowBe4 produces AI translations in 51+ languages for content without an official localization, so a studio-produced module can also reach employees as an AI translation.
- Assignment and recommendations: AIDA Orchestration assigns library training and schedules phishing tests for each user. AI-Selected Phishing Templates pick a template for each user, and AI-Recommended Optional Learning suggests extra modules.
Who Produces KnowBe4’s Core Training Library?
KnowBe4 says its training content “is produced by a studio designed from the ground up to deliver high-production-value training.” According to KnowBe4’s pricing page, the SAT Advanced library adds premium KnowBe4 Studios content such as The Inside Man series, games, and live-action videos.
The studio library treats AI as a subject. KnowBe4 says its catalog adds content on AI-related risks such as deepfakes and shadow AI.
KnowBe4’s public ModStore training catalog also carries third-party material. As of October 2026, its Publisher filter lists KnowBe4 Studios alongside three other publishers: Bisk, APIsec University, and Secure Code Warrior.
Does KnowBe4 Use AI Actors?
KnowBe4 uses AI avatars and AI-generated characters in administrator-built training. Its premium studio library takes a different approach: KnowBe4’s pricing page lists live-action videos and The Inside Man series among the content included in SAT Advanced, and KnowBe4’s library page does not address whether AI actors appear in its studio content, though its help documentation labels some modules “AI-enhanced” for using AI-generated images, video, or voiceovers.
Three KnowBe4 tools add AI actors to custom training.
- Custom AI Video Builder: The builder offers 120+ Synthesia AI avatars for narrated training videos, according to KnowBe4’s July 21, 2026 release.
- Content Creation Agent: AI-generated “cast members” appear in course images and scenarios, according to KnowBe4’s Content Creation Agent Guide.
- Deepfake Training Content Agent: The agent creates a deepfake training video of a real company leader, with a voice cloned from a source video of two minutes or less.
How to Tell Whether a KnowBe4 Module Was Created With AI
KnowBe4 administrators can narrow down which courses and videos came from KnowBe4 AI tools by checking who published each item and where it sits in the KnowBe4 account. Three checks separate content the organization added from publisher-supplied modules.
- Check the publisher: Library modules in the ModStore can be filtered by publisher, and the Publisher filter lists KnowBe4 Studios, Bisk, APIsec University, and Secure Code Warrior.
- Check Custom Content: Courses built with the Content Creation Agent and videos exported from the Custom AI Video Builder are published to Custom Content. Both SAT plans also let organizations upload their own training content, so items in these areas are not necessarily AI-generated.
- Check the Deepfakes subtab: deepfake training videos made with the Deepfake Training Content Agent are listed under Training > Deepfakes and appear in a separate Deepfakes section of the campaign Content menu
Studio modules carry at least one AI exception. KnowBe4 produces AI translations in 51+ languages for content without an official localization, so a module published by KnowBe4 Studios can still reach employees in an AI-translated version.
What AI Features Does KnowBe4 Offer in 2026?
As of October 2026, KnowBe4 offers 12 AI agents under AIDA (AI Defense Agents) plus AI features outside AIDA, including AI-Selected Phishing Templates, AI-Recommended Optional Learning, AI translation, and the Custom AI Video Builder. Some of these features generate new content, such as courses, quizzes, phishing templates, and videos. Others select, schedule, or recommend existing content.
The KnowBe4 AI Feature Map (October 2026) lists each feature, what its AI does, and what it produces, based on KnowBe4’s product pages, support guides, and 2026 releases. The map groups the Orchestration Agent and its three sub-agents in one row.
| KnowBe4 AI feature | What AI does | What it produces |
|---|---|---|
| Orchestration Agent, with the Ongoing Training, Phishing, and Remedial Training sub-agents | Selects, schedules, and generates | Library training assignments and AI-generated phishing tests for each user, plus training tailored to phishing test results |
| Template Generation Agent | Generates | Phishing email templates based on current attack vectors |
| Callback Template Generation Agent | Generates | Phishing templates for the callback attack vector |
| Recommended Landing Pages Agent | Suggests | Landing pages for AIDA-generated phishing templates |
| Knowledge Refresher Agent | Generates and assigns | Bite-sized, scenario-based refreshers |
| Policy Quiz Agent | Generates | Quizzes based on the organization’s security and compliance policies |
| Custom SAPA Agent | Generates | Custom Security Awareness Proficiency Assessments |
| Content Creation Agent | Generates | Text-based courses with a quiz, built from prompts and reference materials and translatable into 51 languages, according to KnowBe4’s Content Creation Agent Guide |
| Deepfake Training Content Agent | Generates | A deepfake training video of a leader from the organization |
| Custom AI Video Builder | Generates | AI-avatar training videos through a Synthesia license |
| AI-Selected Phishing Templates | Selects | A phishing template for each user based on training and phishing history |
| AI-Recommended Optional Learning | Recommends | Optional modules based on previous course completions |
| Simulated Vishing | Generates speech | Voice phishing calls with text-to-speech, local caller ID, and multi-step scenarios |
| AI translation | Translates | Library content in 51+ languages where no official localization exists |
KnowBe4 AI Features by Plan: SAT Foundation vs. SAT Advanced
KnowBe4 launched two security awareness training plans, SAT Foundation and SAT Advanced, in May 2026, and of the two, only SAT Advanced includes AIDA (AI Defense Agents). The plan details below come from KnowBe4’s pricing page, its July 21, 2026 video builder release, and its Content Creation Agent Guide, as of November 2026.
SAT Foundation and SAT Advanced both include these four KnowBe4 AI features.
- AI-Selected Phishing Templates: The system selects a phishing template for each user based on that user’s training and phishing history.
- AI-Recommended Optional Learning: Extra modules are recommended based on previous course completions.
- Risk Score powered by the SmartRisk Engine: Every employee receives a dynamic Risk Score based on behavior.
- Custom AI Video Builder: Administrators create AI-avatar videos through a prepaid Synthesia Starter License.
SAT Advanced adds these three AI capabilities.
- AIDA: Includes all 12 agents, including the Content Creation Agent and the Deepfake Training Content Agent.
- Simulated Vishing and Callback Phishing: Voice phishing simulations with AI text-to-speech, plus callback phishing tests (with AI-generated templates from AIDA), are limited to SAT Advanced.
- AIDA Orchestration Real-Time Coaching Remediation: After a failed phishing test, AIDA chooses between a SecurityTip and a full training assignment.
Three KnowBe4 generators carry published usage limits. The Content Creation Agent allows 25 generated courses per year, reset on the subscription renewal date, and deleting a course does not restore a slot. The Deepfake Training Content Agent is limited to 25 videos. The Synthesia license behind the Custom AI Video Builder covers up to 120 minutes of narrated video per year at no added cost.
How Does KnowBe4 Use AI in Phishing Simulations?
KnowBe4 uses AI phishing simulation features in four ways: choosing a template for each user, generating new templates, automatically scheduling and scaling tests, and generating speech for simulated vishing (voice phishing) calls.
- Template selection: AI-Selected Phishing Templates choose a template for each user based on that user’s training and phishing history in both SAT plans.
- Template generation: The Template Generation and Callback Template Generation agents create templates on demand. Administrators set difficulty with a one-to-five-star rating, which lets AIDA pick the red flags, or they choose the red flags manually.
- Automated testing: AIDA Orchestration generates phishing templates, runs tests automatically, and scales test difficulty based on each user’s performance.
- Voice simulations: Simulated vishing in SAT Advanced uses “advanced text-to-speech,” local caller ID, realistic personas, and multi-step customizable scenarios, according to KnowBe4’s July 30, 2026 release.
What Is KnowBe4 AIDA, and How Does Its AI-Native Training Work?
KnowBe4 AIDA, short for AI Defense Agents, is the set of 12 KnowBe4 artificial intelligence-driven agents included with SAT Advanced. AIDA decides which training and phishing tests each employee receives and generates custom content such as courses, quizzes, and phishing templates. KnowBe4 launched AIDA Orchestration, the agent that automates delivery, on February 24, 2026.
KnowBe4 markets its AIDA and SmartRisk Engine training as AI-native security awareness training. Its product page defines that term as training that “uses AI to personalize content, generate simulations, and deliver coaching automatically, replacing the fixed annual campaign model.” KnowBe4’s timeline says it was “First to bring AI into security awareness training” in 2016 and that its first AIDA patent was issued in 2018.
Based on its product page, KnowBe4’s AI-native training runs as a continuous cycle with three steps.
- Score: The SmartRisk Engine analyzes user behavior and generates a dynamic Risk Score for every employee.
- Assign and test: AIDA Orchestration assigns training from the library and sends AI-generated phishing tests, and KnowBe4 says the SmartRisk Engine helps AIDA prioritize Risk Score.
- Follow-up: After a failed phishing test, AIDA decides whether a quick SecurityTip or full remedial training fits.
KnowBe4 describes administrators’ role in two ways. The AIDA product page says Orchestration works “without the need for human administration.” The AIDA Orchestration Guide says administrators “retain full control” through plans that set included groups, frequency, and excluded content.
How KnowBe4 Personalizes Training With AI and Learning Paths
KnowBe4 says its training is “personalized by AI to each user’s role, risk level, and behavior.” According to KnowBe4’s AIDA Orchestration Guide, the ongoing training agent assigns content from the training library within administrator plans, so for ongoing training, personalization mainly works by selecting library content.
Administrators can exclude content types and topics from a plan, and the plan screen shows the share of the training library AIDA can still assign. AI-Recommended Optional Learning suggests extra modules based on previous course completions, and ModStore Content Recommendations use organization signals such as risk profile.
KnowBe4’s Learning Paths take a different approach to personalization. They organize content into preassembled role- and compliance-specific curricula.
How Does KnowBe4 Rate as an AI Security Training Platform?
As of October 2026, KnowBe4 rates as a broad AI security training platform built on a studio library. It offers 12 AI agents, 1,000+ pieces of KnowBe4 Studios content in SAT Advanced, and automated delivery. Its main limits for buyers are plan gating, published usage caps, and deepfakes delivered as training videos.
KnowBe4’s AI strengths fall into three areas.
- Breadth of agents: KnowBe4 lists 12 AI agents that assign training and generate phishing templates, quizzes, assessments, custom courses, and deepfake training videos.
- Studio library underneath: SAT Advanced includes 1,000+ pieces of KnowBe4 Studios content, and professional linguists and cybersecurity experts localize KnowBe4 content into 47 languages and dialects.
- Automation within plans: AIDA Orchestration sends phishing tests and assigns training automatically, within administrator-defined plans.
Three limits matter for buyers comparing AI security training platforms.
- Plan gating: AIDA, Simulated Vishing, and Callback Phishing are included only in SAT Advanced among KnowBe4’s two current plans.
- Usage caps: The Content Creation Agent allows 25 generated courses per year, and the Custom AI Video Builder license covers up to 120 minutes of narrated video per year. The Deepfake Training Content Agent is limited to 25 videos.
- Deepfakes as training content: KnowBe4 calls its deepfake output “a safe, controlled deepfake training asset.” The asset is deployed in training campaigns or as a standalone module and tracked by completion and engagement.
KnowBe4 fits organizations that want a large studio library, AI-powered delivery, and custom content. Organizations that need simulated attacks built around their own executives and public data should weigh the plan gating and the training-video format of KnowBe4 deepfakes.
AI-Assisted vs. AI-Native Security Awareness Platforms: The Difference
AI-assisted security awareness training is a program that layers AI features, such as recommendations, template generation, and agents, on a human-produced training library. AI-native security awareness training is a program in which AI generates the training content, simulations, and personalization for each organization from that organization’s own data, people, and workflows.
KnowBe4 uses a broader definition of AI-native training: training that “uses AI to personalize content, generate simulations, and deliver coaching automatically.” By that definition, a platform can count as AI-native while the training it assigns comes from a studio-produced library.
KnowBe4 (in SAT Advanced) and Adaptive Security each pair a ready-made library of more than 1,000 items with AI generators. Because both platforms mix the two models, a practical way to compare them is to ask where AI does the work.
Experience the Adaptive platform
Take a free tourAI can do the work in five areas of a security awareness program.
- Content creation: AI writes courses, quizzes, and videos, or the program relies on studio-produced modules.
- Simulation creation: AI writes each phishing, voice, or deepfake simulation, or the program draws on prebuilt templates.
- Personalization: AI decides who receives which existing content, or AI builds the content itself around each organization and employee.
- Deepfake generation: AI recreates a real leader’s face or voice for a training video or a simulated attack.
- Risk scoring: AI scores each employee from behavioral signals and triggers the next training step.

How AI Security Awareness Training Differs From AI-Native Training
AI security awareness training and AI-native security awareness training sound alike but describe two different things. Adaptive Security defines AI security awareness training as training that “educates employees about security risks introduced by artificial intelligence technologies, focusing on recognizing AI-powered attacks and using AI tools safely.”
KnowBe4 uses “AI-native security awareness training” for training that AI personalizes and delivers. A complete program can cover both meanings: AI-related cyberthreats as training topics, and AI as the engine that builds and delivers the training.
Key Benefits of AI-Native Security Awareness Training
AI-native security awareness training offers five main benefits: simulations that match personalized attacks, behavior-based training, organization-specific content on demand, practice across more channels, and fewer manual campaign tasks.
- Simulations that match personalized attacks: In a USENIX Security 2026 field study, researchers generated personalized phishing emails for about $0.03 each, a cost low enough for cyberattackers to personalize lures at scale. Simulations based on an organization’s public data let employees practice responding to those messages.
- Training that responds to behavior: Training and simulations adjust to each employee’s own results. KnowBe4’s AIDA Orchestration scales phishing test difficulty to each user’s performance, and Adaptive Security assigns training based on risk scores and triggered actions. Gartner’s February 2026 cybersecurity trends recommend a shift “from general awareness training to adaptive behavioral and training programs.”
- Organization-specific content on demand: KnowBe4’s Content Creation Agent and Adaptive Security’s AI Content Studio both build training modules from an organization’s own policies. Adaptive Security says AI Content Studio builds a module from a prompt in minutes.
- Practice across more channels: The Verizon 2026 Data Breach Investigations Report (DBIR) found that in phishing simulations, median click rates in mobile-centric vectors such as voice and text messaging were 40% higher than in email. Adaptive Security uses AI to generate email, SMS, and voice simulations from open-source intelligence (OSINT) on each employee, which extends practice to those channels.
- Fewer manual campaign tasks: KnowBe4’s AIDA Orchestration and Adaptive Security’s Auto-Phish both generate and send simulations automatically: Auto-Phish on a cadence the administrator sets, and AIDA Orchestration on a frequency the administrator sets, or AIDA chooses for each user.
What 2026 Research Shows About AI-Personalized Phishing
Independent 2026 research found that AI personalization made phishing more effective against university employees, and its authors recommend awareness training that covers personalized tactics. The evidence comes from a USENIX Security 2026 field study at TU Braunschweig, a German university.
Across 7,741 employees, spear phishing personalized by a large language model (LLM) reached a 10.0% click rate, compared with 3.7% for generic LLM-written phishing and 4.1% for generic human-written phishing. The study measured attack success and did not test training methods.
KnowBe4 AI vs. Adaptive Security: Where Each Platform Applies AI
KnowBe4 uses AI mainly to select and schedule content from its studio library and to generate custom content from administrator prompts and policies. Adaptive Security uses AI to generate simulations and training from each organization’s data, people, and public OSINT data. Both platforms also offer a ready-made library of more than 1,000 items, which KnowBe4 includes in SAT Advanced.
The table below compares the two platforms by where each applies AI, across five areas a buyer can verify, as of October 2026. KnowBe4 details come from its product pages, pricing page, and support guides. Adaptive Security details come from its security awareness training, phishing simulations, and pricing pages.
| Where AI is applied | KnowBe4 | Adaptive Security |
|---|---|---|
| Content creation | The Content Creation Agent builds text-based courses with a quiz from prompts and reference materials, in SAT Advanced, up to 25 courses per year. The Custom AI Video Builder adds AI-avatar video. | AI Content Studio builds modules from prompts, policies, documents, and URLs, auto-personalized with company context, with AI avatar intro videos. |
| Simulation creation | Template Generation agents create templates from administrator settings, and AIDA Orchestration generates and sends tests. | AI generates email, SMS, and voice simulations from OSINT on employees, and Auto-Phish deploys them on a set cadence. |
| Personalization | AIDA Orchestration assigns library training and phishing tests to each user within administrator plans, and AI-Selected Phishing Templates choose a template per user. | Training is assigned by role, dynamic group, department, risk score, or triggered action, and OSINT generates simulations for each employee. |
| Deepfake generation | The Deepfake Training Content Agent creates a training video of a leader from a source video of two minutes or less. It clones the voice and uses a pre-written or custom script. | Custom deepfake personas are built from a photo and a voice sample for use in training. Deepfake simulations cover video, audio, and Teams and Zoom scenarios. |
| Risk scoring | The SmartRisk Engine Risk Score combines training completion, phishing results, and coaching response in both SAT plans. | Per-employee risk scores update from simulation outcomes, training completion, reported phishing, and other signals. |
For organizations that want AI to build simulations around their own executives, public data, and channels, Adaptive Security fits. For organizations that want to keep KnowBe4’s studio library and automate its delivery, KnowBe4 SAT Advanced with AIDA fits.
Which Is Better for AI-Powered Phishing Simulations: Adaptive Security or KnowBe4?
Adaptive Security is the better fit for an AI phishing simulation program that needs SMS, voice, and deepfake attacks generated from open-source intelligence (OSINT) on each employee. KnowBe4 fits programs that want AI-generated email, callback, and voice phishing simulations within its SAT Advanced plan.
Three criteria decide which platform produces more realistic AI-powered phishing simulations.
- Channels: Adaptive Security phishing simulations cover email, SMS, voice, deepfake video and audio, and multi-channel attack chains. KnowBe4’s pricing page lists AI-Selected Phishing Templates, Callback Phishing, and Simulated Vishing.
- Source of personalization: Adaptive Security generates simulations from OSINT data on employees. KnowBe4 generates templates from administrator settings and selects templates from each user’s training and phishing history. By default, AIDA can also use real-time coaching events from connected security tools to personalize phishing tests.
- Plan and add-on terms: Adaptive Security prices voice phishing as a separate add-on, and KnowBe4 includes Callback Phishing and Simulated Vishing only in SAT Advanced.
Which Security Awareness Platforms Use AI to Create Custom Training?
Security awareness platforms that use AI to create custom training include KnowBe4 and Adaptive Security. KnowBe4 uses its Content Creation Agent and Custom AI Video Builder, and Adaptive Security uses AI Content Studio and AI-generated simulations. The two platforms take different approaches to custom content.
- Library-first platforms with AI generators: A studio-produced library remains the core, and administrators use AI tools to add organization-specific courses and videos. KnowBe4 follows this approach and describes its AI tools as a way to turn organization-specific policies into lessons alongside its “handcrafted, studio-grade content.”
- Platforms that generate content from the organization’s own data: AI builds simulations and training around the organization’s executives, public data, and policies. Adaptive Security follows this approach with simulations generated from OSINT data and modules auto-personalized with company context, alongside a library of 1,000+ interactive modules.
Adaptive Security is a security awareness training and phishing simulation platform that creates deepfake video and audio, voice, SMS, and email simulations built from public OSINT and each employee’s role, with optional custom executive deepfake personas. For security teams choosing an AI security training platform that builds training around their own policies and leaders, Adaptive Security provides AI Content Studio, which generates modules from prompts or policies.
Adaptive Security also creates custom deepfake personas from a photo and a voice sample, which lets employees practice responding to impersonations of their own leaders. Adaptive Security offers training in 40 languages with dubbed narration and prices voice phishing as an optional add-on.
The AI Readiness Checklist for Security Awareness Platforms
The AI Readiness Checklist for Security Awareness Platforms gives buyers seven questions to ask any vendor in a demo, each tied to a fact the vendor can document.
- Where does AI generate content, and where does it only select? Buyers should ask the vendor to separate generated assets, such as courses, simulations, videos, and translations, from library assets that AI only assigns. For example, KnowBe4 pairs a studio-produced library with AI generators and AIDA Orchestration, which assigns library training.
- Who reviews AI output before employees see it? Buyers should ask which review points exist, such as outline approval, full editing, and a learner preview. KnowBe4’s Content Creation Agent and Adaptive Security’s AI Content Studio both show an outline before generating a module.
- Is customer or employee data used to train or fine-tune models? Buyers should ask for the commitment in writing. The Federal Trade Commission (FTC) has warned AI model-as-a-service companies that breaking promises not to use customer data for model training can lead to FTC enforcement.
- Which AI features are in the quoted plan, and which limits apply? Buyers should ask for plan-level detail and usage caps. KnowBe4 includes AIDA with SAT Advanced and caps generated courses at 25 per year, and Adaptive Security prices voice phishing as a separate add-on.
- Can simulations reproduce personalized, multi-channel attacks? Buyers should ask which channels the platform simulates, such as email, SMS, voice, and video, and which data personalizes each simulation. KnowBe4 simulates voice phishing with text-to-speech in SAT Advanced, and Adaptive Security generates email, SMS, and voice simulations from OSINT on employees.
- Which authorization controls govern deepfakes of real executives? Buyers should ask how the vendor confirms consent and where it stores source media. KnowBe4 requires “legal authorization from the individual being recorded,” and Adaptive Security’s terms require customers to hold “all necessary rights, permissions, and consents.” The NIST Generative AI Profile (July 2024) lists unauthorized use or disclosure of biometric data among generative AI privacy risks.
- How does the platform measure behavior change beyond completion? Buyers should ask which behaviors feed the risk score and how trends are reported. NIST SP 800-50 Rev. 1 (September 2024) addresses measuring the impact of learning programs, and both KnowBe4 and Adaptive Security score each employee from behavioral signals.
What Are the Risks and Limits of AI-Generated Security Training?
AI-generated security training carries three main risks: inaccurate content that reaches employees without review and loss of control over brand and executive likeness. Review steps and authorization controls address each risk.
Why AI-Generated Training Needs a Review Step
AI-generated training needs a human review step before it reaches employees because generative AI can state false information confidently. The NIST Generative AI Profile (NIST AI 600-1, July 2024) calls this risk confabulation and recommends evaluating generative AI output accuracy using methods that include human oversight.
KnowBe4 and Adaptive Security build that review into their generators. KnowBe4’s Content Creation Agent shows a course outline before generating content, and Adaptive Security’s AI Content Studio adds an adjustable outline that administrators review before generation. Adaptive Security’s March 2026 release states the principle directly: “The best AI output still needs a human finish.”
Brand Control and Executive Likeness in AI-Generated Training
AI-generated training places an organization’s brand and its leaders’ faces and voices in software-produced content. KnowBe4 requires “legal authorization from the individual being recorded” before creating a deepfake of a leader and documents the resulting videos as educational training content.
KnowBe4’s Deepfakes Training Guide also notes that the deepfake feature falls outside the scope of KnowBe4’s Federal Risk and Authorization Management Program (FedRAMP) Moderate authorization. Organizations that rely on that authorization should account for the gap before deploying deepfake training.
Adaptive Security’s terms of service require customers to warrant that they hold “all necessary rights, permissions, and consents” from people whose images or voices appear in AI-generated content. The terms also bar using that content to impersonate anyone outside authorized training, testing, or educational scenarios.
Both vendors offer brand controls for generated training. KnowBe4’s Content Creation Agent uses themes to set each custom course’s visual identity, and Adaptive Security customizes modules with the organization’s logo, colors, fonts, imagery, and email domain.
The Future of AI in Security Awareness Training: 2026 Trends
Security awareness training is moving toward behavioral programs that use AI to generate content and simulations, driven by AI-enabled cyberattacks and by employees’ own use of AI. Gartner’s 2026 cybersecurity trends state that “Existing security awareness efforts continue to fail to reduce cybersecurity risks as GenAI adoption accelerates.”
Four trends are shaping AI in security awareness training in 2026.
- Vendors expanded AI content generators: Adaptive Security introduced a completely revamped AI Content Studio in March 2026, KnowBe4 launched its Content Creation Agent in May 2026, and KnowBe4 added the Custom AI Video Builder in July 2026.
- Cyberattackers use AI across many techniques: The Verizon 2026 Data Breach Investigations Report (DBIR), based on more than 22,000 confirmed breaches, found that the median threat actor researched or used AI assistance in 15 different documented techniques.
- AI use became a training topic: KnowBe4 says its catalog adds content on AI-related risks such as deepfakes and shadow AI, and Gartner recommends programs “that include AI-specific tasks.”
- Vendors now position their products around AI agents: KnowBe4 describes itself as “securing both humans and AI agents,” and Adaptive Security positions its products as “Human and agent security for the AI era.”
Frequently Asked Questions About KnowBe4 AI
Is AI-Generated Security Training Accurate?
AI-generated security training is only as accurate as its source material and the review it receives before publishing. KnowBe4’s Content Creation Agent accepts reference materials alongside the prompt and requires administrators to accept a legal disclaimer before it generates a course. Translation is a second accuracy point. Professional linguists and cybersecurity experts localize KnowBe4 content into 47 languages and dialects, while content without an official localization receives AI translation in 51+ languages.
Do Administrators Review AI-Generated KnowBe4 Content Before Release?
Administrators can review AI-generated KnowBe4 courses and phishing templates before employees see them. The Content Creation Agent opens each finished course in an editor where administrators can change text in any frame and preview it as learners will see it. Generated phishing templates also open for viewing and editing. AIDA Orchestration automatically sends phishing tests, so administrators do not approve each test in advance. They control tests through each plan’s groups, frequency, and exclusions, and can review sent tests and AIDA’s reasoning in the Phishing Activity report.
What Data Does KnowBe4 Use to Personalize Training?
KnowBe4 personalizes training and phishing tests based on each employee’s Risk Score, which combines training completion, phishing simulation results, and coaching responses. By default, AIDA can also draw on real-time coaching events from connected security tools, such as visits to risky websites. Administrators who want to limit that data can turn on the plan setting “Exclude Real-Time Coaching activity from phishing test personalization,” and AIDA then personalizes tests from the user’s phishing and training history.
Can Administrators Pause or Turn Off KnowBe4’s AI Agents?
Yes. Administrators can switch AIDA Orchestration on or off with the Enable AIDA Orchestration toggle on the Dashboard. Within each plan, they can turn the phishing agent and the ongoing training agent on or off. They can also deactivate AIDA remedial training, which switches that agent to Test Mode. Content generators such as the Content Creation Agent and the Deepfake Training Content Agent start from an administrator’s prompt or uploaded source video.
Can Organizations Upload Their Own Training to KnowBe4?
Yes. Both SAT Foundation and SAT Advanced include an “Upload Your Own Training Content” feature for managing an organization’s custom training in its KnowBe4 account. Organizations on SAT Advanced can also generate custom courses with the Content Creation Agent, then publish them to Custom Content or download them as Sharable Content Object Reference Model (SCORM) packages for an external learning management system.
Choosing Between AI-Assisted and AI-Native Security Awareness Training
Choosing between AI-assisted and AI-native security awareness training means deciding where AI should do the work: on top of a studio-produced library, or in content built around the organization. The AI Readiness Checklist for Security Awareness Platforms gives buyers seven questions to test any vendor.
As of October 2026, KnowBe4 describes its core training as studio-produced, and KnowBe4 AI features add custom courses, phishing templates, videos, and automated delivery, mostly in SAT Advanced. Adaptive Security uses AI to generate simulations and training around each organization’s own executives, policies, and channels. Security teams that want to see that approach in practice can take a self-guided tour of Adaptive Security.
As experts in cybersecurity insights and AI threat analysis, the Adaptive Security Team is sharing its expertise with organizations.
Get started with Adaptive Security



