These AI glasses promised to make me smarter, and all they got was Clippy for my face

This is it Mohsena weekly newsletter sent every Friday from one of the Verge’s top reviewers Victoria song Which analyzes and discusses the latest phones, smartwatches, apps and other gadgets that you swear will change your life. Mohsen It arrives in our subscribers’ inboxes at 10 a.m. ET. Subscribe to Mohsen here. We will be taking a two week hiatus and will be back on November 7th.

As I am Written last weekMy body is quickly running out of parts to do my job. Part of being human is knowing when to ask for help, so a few months ago, I recruited senior editor Sean Hollister — a fellow smartglasses geek — to help me test Halo Glass, an always-listening AI companion that lives inside a pair of glasses.

Halo is the brainchild of two former Harvard students who made headlines last year afterward He rigged a pair of Ray-Ban Metas to do stranger things in real time. In August, AnhPhu Nguyen and Caine Ardayfio Announce They’ve been making a pair of always-on AI glasses that can listen to conversations, record them, transcribe them, and then organically feed you answers to questions relevant to your conversations in real time. It’s kind of a mix in between Clearlyanother AI startup that aims to help you “cheat at everything.” beean AI wearable that claims to be your second memory. Instead of a pin or wristband, this allows you to discreetly view answers inside a pair of smart glasses.

So of course I wanted to test them.

Sean and I spoke with Ardayfio, who told us that although Halo will eventually make its own hardware, we’ll be among the first to try running their app on Even G1 reality glasses. You may not have heard of Even Realities, but they are among the most admired smart glasses makers At CES. All we have to do is try out the prototype, compare notes, and then write up our experience. Easy, right?

The appeal for both of us was to have a second memory. We are busy, and sometimes we forget. Wouldn’t life — and our jobs — be a little easier if we stopped forgetting that one thing we told our colleagues, bosses, and spouses we would do? Wouldn’t interviewing sources be easier if the definition could emerge in real time, when they used an esoteric term, without having to interrupt the flow of the conversation?

Sure, it sounds nice, but AI-powered wearables always present a boatload of ethical conundrums. Since this is a fairly new product category, the ethics surrounding it surprised us a little. For starters, Sean lives in California, a state that legally requires both parties to agree to a recording of the conversation. Would he commit a crime if he wore these glasses without revealing to everyone around him that he was recording? Sean’s wife has a job that requires secrecy. The always-on recording device could jeopardize her livelihood if Shawn forgets to turn it off while she’s working and he’s nearby. As a result, Sean I cannot Actually test these glasses at home. Meanwhile, my wife is tired of AI wearables that always listen after I review bee And he copied one of our fights. (For testing friend(I had to wear them outside the house.) The solution we came up with was for each of us to put on a pair of Halo-powered G1 goggles and do a video call to test them out with each other.

In theory, the Halo app works like this: In the app, you see a live version of the conversation happening around you. Every now and then, a pop-up appears about something you’ve mentioned. For example, maybe you’re talking about animals native to Australia, and someone asks you which is the most dangerous. This answer is sent to your glasses, and you can sound like a smart pantsuit in your conversation. Once the conversation is over, you see a quick summary of it and some action items to address — similar to what you gather at the end of a meeting.

In practical terms, our invitation was ridiculous.

I started with a 20-minute troubleshooting session that included multiple firmware updates and disconnections. I’ll spare you the details except for these, because they’re just… the most awkward way imaginable to interact with the AI: To summon the screen, the G1 glasses require you to look up. You can adjust the desired angle — a wise choice, since the default value is 40 degrees. This is like just throwing your head back to look at the ceiling. We’ve both adjusted to about 15 degrees, but that’s still a comically obvious trigger.

Wonky prototype hardware can be forgiven because you’re exploring an idea. And the idea that AI glasses could make you seem smarter without knowing who you’re talking to makes me uncomfortable.

I spoke to Sean about my concerns. We debated whether smart glasses really help people stay present in the moment. We wondered, can you really be yourself if you know you’re being recorded? What is the level of ethical disclosure? How do you protect the privacy of your loved ones who may not be as keen on this technology as you are?

It was an interesting conversation, except for all the times the AI ​​got involved. At that point, one of us had to throw our heads back to see what alert had appeared. Imagine Sean and I, 30 minutes into the call, throwing our heads back like deranged sea lions barking on the dock.

He may sometimes interfere in trivial, useless matters. For example, the definition of “hidden” came up for me after I used it correctly. I was slightly offended that the AI ​​probably thought I didn’t know what that word meant in context. When I pointed out ClearlyInstead, Halo AI provided facts about it ignorant“, “A 1995 coming-of-age comedy directed by Amy Heckerling.” Paradigmatic Artificial Intelligence.

The worst was when Halo showed a message explaining that cell phones first appeared in the 1970s and 1980s. Sean must have said something to me about phones for me to receive this message. I relayed the truth to Shawn. Then he told me his glasses showed the same notification. AI once again alerted me that phones first appeared in the 1970s and 1980s. We were stuck in a hellish AI-powered Ouroboros game. We shook our heads some more.

Halo AI has provided useful facts several times. The definition of “nits” came up when we were talking about displays on smart glasses. I defined “annihilation” when Sean and I were thinking about the implications that permanent recording would have on the lives of the people around us.

But ultimately, using Halo was more of a distraction than a help. All the while, roughly 10% of brainpower was spent wondering when the assistant would intervene or disconnect. When rereading the transcript of our conversation in the app, there were a lot of omitted topics that I wished we had delved into, if not for all the distractions.

Sean told me that his interest in Halo was due to a very human desire to “remember better.” I bet anyone with a to-do list would do this. I felt the same way when testing the AI ​​wearable Bee. And after, this Conversation — where the AI ​​was broadcasting the same facts to each of us at the same time — just reminded me of Microsoft’s Clippy. Always there, pestering you with tidbits that weren’t exactly helpful, interrupting your train of thought just as you went along.

For now, I think I’ll accept my imperfect mixture of analogue note lists and to-do lists. I’ll just look stupid in the conversation by asking, “I’m sorry, what does that mean?” It’s not sexy, but I’d rather not shake my head next time I need an answer.