35 Years Ago, Star Trek Made A Chilling AI Prediction — With A Twist
When will we tell our AI not to tell us stuff?
Even quiet, and somewhat under-discussed episodes of Star Trek: The Next Generation have oddly prescient qualities. During its nearly-perfect fourth season, TNG had several bangers, and also some episodes that flew under-the-sensors. During the week of February 11, 1991, one of those episodes was “Clues,” the 14th episode of Season 4, which, at the time, seemed like a fun head-scratcher. But today, the episode scans somewhat differently than it did back in 1991. What was then a strange story about a ship-wide mystery actually now reads as a kind of predictive tale that applies to 21st-century anxieties about what AI does, and doesn’t tell us about the truth.
Here’s why “Clues” is worth a rewatch now, and how Mr. Data (Brent Spiner) acts as an aspirational AI, with a big twist.
Spoilers ahead.
In the 2014 movie Interstellar, a running joke is made that the robot TARS is always lying to the crew a little bit, otherwise the humans would get uncomfortable. This idea harkens back to some of the most formative speculative fiction about artificial intelligence; we want our robots to serve us like tools, but we also don’t always like the answers generated by databases and algorithms. In the 1941 Isaac Asimov story “Liar!” a robot starts to lie to his various human colleagues, out of concern for their feelings, and to slavishly adhere to his do-no-harm programming, it begins telling white lies.
Hold the phone! Our AI is lying!
Pitched by a fan named Bruce D. Arthurs, the spec script for “Clues” was later rewritten by TNG staffer Joe Menosky. This open-door policy, that TNG would take script submissions from outside their writers’ room, began in Season 3, and represents one of the coolest things about the iconic series; the moment when all ideas were welcome, and could become legit episodes of Star Trek.
Appropriately then, the concept of the TNG episode “Clues” is about insider knowledge versus what we can call “the truth.” At first, the episode starts with a space mystery: the Enterprise slips into a wormhole and loses an entire day. Everyone passes out, and Data, an android, is the only person not affected. As the episode goes on, various details indicate that something has happened that doesn’t add up with Data’s assessment of what’s happened. After dropping some obscure technical mumbo-jumbo to hand-wave away the explanation for why everyone was knocked out, Captain Picard (Patrick Stewart) straight-up asks the crew: Is Data, their infallible AI friend and confidant, lying to all of them?
The episode then takes on a double mystery: What happened to the crew, and why would Data lie about it? Like the robots of old, Data is programmed with Asimovian laws; he has to obey orders from humans, but he also can’t allow people to come to harm through action or inaction. So, like many good Star Trek conflicts, the question here isn’t about malevolence or deceit with evil intention, but instead, about some more complex problem. Data hasn’t turned bad and taken control of the ship (that’s for a different episode), but he is lying. Why?
Comparing Data to modern-day AI chatbots is tempting, though it is a slippery analogy. ChatGPT and Grok aren’t sentient, nor do they have a desire to become more human, take tap-dancing lessons, or fall in love. Data is, paradoxically, an analog AI; he’s not part of a cloud of shared information like the Borg or the internet, and his decisions run on hardware that isn’t “online,” so to speak. And yet, in the final moments of this episode, we learn that Data was ordered to lie to the crew by Picard himself. The reason? A very dangerous and isolationist alien species demanded to be kept secret, and the only solution was a mass memory wipe, and for Data to be ordered to keep the secret forever.
Can we trust Data? Can we trust AI?
In today’s context, this isn’t really about AI hallucinating or giving false information, but instead, hiding what has been deemed harmful information from the population at large. Does this happen today? You bet. You can download apps that prevent you from going online, content filters exist everywhere, and various chatbots are prevented from talking about specific things, for somewhat obvious reasons.
In the Star Trek prediction of our AI future, though, the organic folks would still be in complete control. The average person doesn’t have a good sense of what AI is and isn’t right about, making all of us more like the unnamed crew members of the Enterprise who never heard about this scandal, and never knew Data had been told to lie. In our minds, we’re all like Captain Picard, telling our technology what to do and when to do it. But, in practice, the future of AI deception — intentional or not — is murkier. The Enterprise crew sleeps easily in the end, knowing that if Data is lying about a bunch of stuff, they can still trust him. It’s too bad that in our brave, new world, Mr. Data is still an utterly anachronistic kind of artificial life form. If we knew Data was fact-checking ChatGPT or Grok, we all might feel better too.
