Imagine a finance employee gets a video call from someone who looks and sounds exactly like the company’s CFO, asking for an urgent wire transfer. If the face and voice were generated by AI, that call is a deepfake. The same techniques can put a real person’s face on someone else’s body in a video, clone a voice from a few minutes of audio, or create a photo of an event that never happened.
The name comes from “deep learning,” the family of neural-network methods used to make the fakes. Lawmakers now define the term directly. The EU AI Act calls a deep fake AI-generated or manipulated image, audio or video content that resembles real people, objects, places or events and “would falsely appear to a person to be authentic or truthful” (EU AI Act, Article 3(60)). The definition does not require bad intent. Satire, film effects and consented voice clones can be deepfakes too; what makes one harmful is using it to deceive.
Hiring is one place the harm is concrete. The FBI’s Internet Crime Complaint Center warned in 2022 about applicants using deepfake video and voice, combined with stolen personal data, in interviews for remote IT jobs that would give access to company systems (FBI IC3). On the show, Greenhouse CEO Daniel Chait describes the same pattern in episode 58 at 16:01: an “industrial scale” effort to infiltrate remote IT jobs using deepfakes, with a call center of people taking turns posing as one candidate. He links it to the flood of AI slop applications that makes fakes harder to spot. Episode 58 is sponsored by Galileo, the Season 3 presenting sponsor; the read is at 2:25.
It matters because seeing and hearing someone used to be decent proof of identity, and that is no longer safe to assume. Detection tools exist, but they are not reliable enough to be the only defense. Our practical summary: verify high-stakes requests through a second channel you already trust, such as calling back on a known number, and treat identity checks for remote hiring and payments as a security problem rather than a judgment call made on one video.
Hear it from the guest
“There's a growing industrial scale North Korean effort to infiltrate remote IT jobs using techniques like deepfakes”
Quotes lightly edited to remove filler words.
Sources
- EU AI Act, Article 3(60): definition of deep fake — Defines a deep fake as AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful; checked October 24, 2026.
- FBI IC3: Deepfakes and stolen PII used to apply for remote work positions — June 2022 public service announcement reporting complaints of deepfake video and voice used in remote job interviews, and describing deepfakes as video, images or recordings convincingly altered to misrepresent someone; checked October 24, 2026.
- Chain of Thought, episode 58: Daniel Chait on deepfake job candidates — Chait describes organized efforts to infiltrate remote IT jobs with deepfakes and shared fake identities.
