Today’s cybersecurity pros have their own kind of Voight-Kampff test as they try and figure out: “Are you a human or not?”

Morning Brew Inc, Photos: Warner Brothers
Early on in the 1982 sci-fi thriller Blade Runner, agent Rick Deckard (Harrison Ford) puts a bioengineered, human-like being known as a “replicant” to the test—specifically, a “Voight-Kampff” test. The tense exam measures interviewee responses to a series of odd or emotionally provocative questions and statements.
For example, Deckard asks: “You’re watching television. Suddenly you notice a wasp crawling on your arm.”
Careful with your answer. In the movie, your answer determines if you’re human, which means Deckard will let you live.
While today’s cybersecurity pros don’t have time to sit across from every bot and ask it ethical questions, there are plenty of real-life, modern security mechanisms (like a CAPTCHA) to determine a human visitor from a fake one.
Like Deckard facing a new generation of even more human-y replicants, the job of sorting bots from humans has only gotten tougher thanks to the rise of AI agents and sophisticated deepfakes. That job applicant on a video call, for example, could be a digital fake using an AI-generated knowledge base to answer questions.
“40 years ago, that was science fiction. That is no longer science fiction for us,” Rich Greene, certified instructor at cybersecurity training provider SANS Institute, told IT Brew. “That really is the cybersecurity problem that we’re actually trying to solve right now.”
Zero trust. The Blade Runner replicants’ capabilities improve and diversify, from the acrobatic flash of Pris (Darryl Hannah) to the wise, reflective nature of Roy Batty (Rutger Hauer). Human-like AI has been getting better, too:
Enter today’s “replicant hunter,” the cybersecurity professional tasked with spotting the bots—not with pupil-dilation testers or emotional questions, but with less cinematic tools like multi-factor authentication, behavioral analytics, and log analysis.
“We are constantly asking whether a login, an email, or even a phone call or a Zoom call is really coming from who it claims to be,” Greene said.
Funnily enough, Greene’s advice to today’s security pros almost sounds like the opposite of a Voight-Kampff test: Don’t waste time analyzing the deepfake. Rather than closely examining a video or email message, pick up the phone and ask the sender a question—one more relevant than, “Do you kill the wasp?”
“Take the actual moniker of zero trust and put it in real life, not just digital technology,” he said.
Well, some trust. People today want a level of humanity from their AI, and they increasingly appear to trust the tech with more personal communications. A Harvard Business Review study of over 12,000 use cases, released in June, found that, for the second year in a row, the top AI use was therapy and companionship.
So, will human-like AI be our companion or our adversary?
Greene sees exciting AI possibilities in the medical and financial fields, and has faith in the cybersecurity industry to adapt to AI’s disruptive changes, as it did with the emergence of mobile computing and the cloud. Also, there’s one capability that today’s AI still can’t quite replicate, he told us: intuition.
If AI hits a roadblock, “it will make up some completely nonsense, hallucinating garbage,” Greene said. “Because it doesn’t have the ability to truly think on its own, or create from nothing.”
We may have awhile to go, in other words, before we reach Roy Batty giving a soulful soliloquy.
For more IT Brew movie club, check out our dives into Hackers, Sneakers, Blackhat, Skyfall, The Matrix, Jurassic Park, and WarGames.