I recently reviewed Even Realities G1 smart glasses (they’re very cool) and the first real-world thing I used them for was scamming someone. I told a local bartender I had an encyclopedic knowledge of film, and I would answer the hardest movie trivia question he could come up with in exchange for a drink. After a short consultation with Google, dude came back with “Who directed 1922’s Cabinet of Dr. Caligari?”
I tilted my head thoughtfully and repeated the question as if verifying I heard him right. The AI agent silently did its thing, and in about three seconds, the answer was floating before my eyes, totally invisible to everyone around me.
Credit: Stephen Johnson
“Robert Wiene?” I asked, feigning uncertainty. Boom! Free drink. It’s not the hardest trivia question, but I could have answered literally anything: the date Dr. Caligari was released, the day of the week it was, or the weather that day—all short work.
Bartenders know to be wary of bar bets, so this one was watching me like a hawk to make sure I wasn’t looking up the answer on my phone or something. He didn’t seem to suspect my glasses, and even if he had, it wouldn’t have mattered. Even inspected up close, G1s offer no indication of electronics of any kind: No USB port, no flashing light, no visible controls.
The AI agent is activated with a subtle tap behind the ear. Repeating the question sends the AI off for the answer, presented in a display that’s invisible to everyone but the wearer. You could do the same trick with a pair of Meta Ray-Ban Display glasses or audio-only smart glasses with AI.
I don’t like deceiving people, so I ended up telling the bartender what was up to and not accepting the drink, but it got me thinking about what more nefarious people than I could do (and probably already are) with AI smart glasses.
Just imagine what I did on a bigger level: a team of hustlers at bar trivia silently tapping their temples whenever “which actor played…?” was asked would never lose. It’s pretty small stakes, but not hard to imagine. Hypothetically, a hacked pair of smart glasses could be programmed to read the cards in your poker hand and give you the probability of winning in real time, either through the display or whispered in your ear. They could, hypothetically, make counting cards in blackjack effortless and undetectable. Stretching it out further, hypothetically, glasses could scan others players for “tells” that they’re bluffing, or read micro-expressions to give a constantly updated read on opponents.
Along the same lines, imagine attending a self-help meeting, and the leader, who you have never met, says, “I’ve had a vision about you” and begins to describe something that happened in your life exactly. All it would take would be glasses that recognize your face (supposedly coming to Meta Displays) connecting to a social media feed, which is then displayed in real time to the leader. Or you could go more subtle and engineer small “serendipities” like mentioning a movie someone recently saw and having exactly their opinion. It would only take a few of these and maybe some mystical patter to convince people you are divine being they should definitely donate to.
How to spot tricky smart glasses
I could go on, but you get the idea. Luckily, there are some indications when someone is using tech hidden in their eyeglasses. The most important is a general understanding of the possibilities of this technology. If someone seems to know something they shouldn’t, ask yourself if their glasses could be the source of their power. Here are some more specific giveaways.
Look for identifiable kinds of smart glasses
Credit: Meta
The most popular display style glasses, Meta Ray-Ban Displays, are distinctive looking, with a fairly obvious camera in one of the corners and a specific look and branding. But other kinds of smart glasses, like the Even Realities G1s mentioned above, are obscure enough that most won’t recognize them, and so “normal” looking that most people wouldn’t pick them out of lineup.
Look for where the tech is hidden
Most smart glasses are still fairly bulky, so look for thick arms or frames where the wires are hidden. But again, that’s only most smart glasses; some are totally sleek, with imperceptible tech.
What do you think so far?
Look for a small glint
In most situations, the display in display glasses is not visible to anyone but the glasses wearer, but there’s still light being projected. In a dark room, you can see a green glow, and even if it’s not dark, the display windows are visible if the light hits it just right, as you can see here:
Credit: Stephen Johnson
But honestly, it’s subtle and hard to spot.
Listen for the sounds
Older styles of audio-only smart glasses can feed information to wearers, but the open-air speakers mean some sound is bleeding into the atmosphere. You can definitely hear smart glasses if you’re in a quiet room and you’re trying to. If it’s loud, operating glasses becomes problematic for the wearer.
Look for the source of control
Smart glasses have to be controlled somehow. Meta Display glasses are operated with a wrist band. G1 glasses’ AI agent is powered on by tapping on the frames behind your ear. Anyone who practices for a few hours could make these movements seem natural, but they’re there if you know what to look for.
Odd movements and speech
It doesn’t take advanced stagecraft to operate these kinds of glasses imperceptibly, but it does take something. Tells might include small glances upward to see the display, tapping on glasses, stilted speech while waiting for information to come in, and a reading-a-teleprompter style of talking. Watch for people repeating questions back. But understand, it’s hard to spot. When I was conning my bartender, I thought how I repeated the question back was obvious, but my wife said I just seemed a little odd, which suits the personality of a trivia whizz anyway.
Smart glasses are powerful tools, like a hammer or a calculator. Like any innocent tool, they can be used for nefarious things, so until manufacturers or regulators require obvious indicators, like flashing lights or visible controls, we have to protect ourselves by paying attention to these small cues and staying skeptical when someone seems to know more than they reasonably should.