KnowBe4 shows employees an admin-generated deepfake video of an executive. Kinds has each learner clone their own voice, hands-on. Two different ways to teach AI threats.
The short version: Both platforms teach employees about deepfakes and voice cloning, but the learner's role is completely different. In KnowBe4, an admin uploads a sample of a company executive and the platform generates a deepfake video of that leader for employees to watch. In Kinds, the learner records a short sample and hears an AI clone of their own voice inside the workshop. One is watched. One is experienced.
Does KnowBe4 train employees on deepfakes and voice cloning?
Yes. KnowBe4 offers an AI Voice Cloning Demo video module (available on its legacy Silver, Gold, and Platinum tiers) and, more notably, the AIDA Deepfake Training Content Agent, launched around December 2025 on SAT Advanced. An admin uploads a short video or audio sample of a company executive, and the platform generates a deepfake video of that leader for training.
Seeing a convincing fake of your own CEO is a strong hook, and KnowBe4 deserves credit for shipping it. The company also partners with Synthesia for AI-avatar training videos in 130+ languages and offers a Content Creation Agent that turns policies and text into training. KnowBe4's product-side AI runs deep too. It says it has used machine learning since 2016 and trademarked AIDA in 2017.
Is KnowBe4's deepfake training hands-on for the learner?
No. The deepfake is created by an admin, it is about a company leader, and employees experience it as a video to watch. The learner never touches the cloning technology themselves. It is a demonstration, produced centrally and consumed passively, like the rest of a video-first library.
That matters because the lesson of voice cloning is not "fakes exist." It is "a fake of anyone can be made in minutes from a small sample." A demonstration shows the first. It cannot make someone feel the second.
How does Kinds Security teach AI threats?
Inside a Kinds workshop, the learner records a few seconds of their own voice and then hears an AI clone of it speak. No admin sets anything up. The point lands personally: if this took thirty seconds with my voice, it works on my voice, my boss's voice, and my family's.
The workshop then walks through what to do about it, in the same chat-style, under-7-minute format as the rest of Kinds' training. Nobody has to source executive footage, get sign-off to deepfake the CEO, or schedule a rollout. Every learner gets the experience the moment the workshop is assigned.
How do the two approaches compare?
Kinds Security | KnowBe4 | |
|---|---|---|
Who experiences the AI | The learner, hands-on, with their own voice | Employees watch a video an admin generated |
Voice cloning | Learner clones their own voice inside the workshop | AI Voice Cloning Demo video module |
Deepfake video | Not the focus; training centers on voice | Deepfake Training Content Agent makes a video of an executive (SAT Advanced, launched around December 2025) |
Setup required | None; built into the workshop | Admin uploads executive video or audio and distributes the result |
Languages | English-first | Synthesia AI avatars in 130+ languages |
Bottom line
If you want a dramatic, centrally produced deepfake of your own leadership, and you are on SAT Advanced, KnowBe4 ships that today. If you want every employee to personally experience how fast their own voice can be cloned, with zero admin production work, that is what Kinds built. You can try it free for 21 days, no demo call.
FAQ
Is KnowBe4's deepfake agent available on every plan? No. It sits in SAT Advanced, the top tier after KnowBe4's May 2026 repackaging (Diamond plus AIDA under the old names).
Does the Kinds voice clone leave the platform? The clone exists to teach the learner during the workshop. It is a training exercise, not a media asset an admin distributes.
Which approach is safer to roll out? Deepfaking an executive usually needs legal and comms sign-off. A learner cloning their own voice needs nobody's permission but their own.
Do both cover AI phishing more broadly? Yes. KnowBe4 covers AI threats across its library and agents. Kinds bakes AI-threat scenarios into its interactive workshops.
