Omi AI is Not Your Friend

Have you ever wondered what it would be like to have a perfect memory? Never forget a name, always have the context for that important meeting on the tip of your tongue; be able to know exactly why an argument unfolded the way it did?

That's the promise of a new range of wearable artificial intelligence products: as long as you're willing to sell your soul for the privilege. I've explored each of these devices in depth, and what I've discovered haunts me to my core.

Tape recorders have been around since forever, helping people record memos. Human memory is fallible, and we're deeply bad at recalling truth. What you think happened during that argument with your wife or husband? Probably not what actually occurred. You color your experiences with emotion, with tunnel vision, and with the context of your internal dialogue.

But what if an AI tool that you wore around your neck could change all of that? What if, in the middle of an argument, your AI app popped up a message that read: "Hey, you actually said this, and you should take a deep breath and then apologize."

As the world heats up and our lives degrade beneath the feet of billionaires, technological tools will become normalized to the point of dependency.

Tools are not evil. For all that the AI industry is cooking the planet, the technology itself is just another stepping stone in a long line of historical memory aides, starting with the capacity for speech itself.

But as with any tool, the way it gets used can determine whether the outcome is good or evil.

Imagine being recorded without your consent.

Imagine having your most intimate conversations transcribed on a server that anyone with basic computer skills can access.

Imagine a world where your every moment is surveilled by the powers that control your access to healthcare, to home insurance, to your driver's license.

Omi is an open-source AI nightmare

You may have heard their names already: Plaude. Limitless. Humane Ai pin. And that's not getting into the glasses produced Zuckerberg's terrifying Meta empire and similar behemoths.

I was told that enhancing the privacy of the app conflicted with the payment model of the company

I'm fascinated by these tools and the benefits they could lend to my life. But as I began to explore them, I encountered a horror show of privacy failures at every turn. Even options like the Limitless Pendant, where a relatively clear focus on privacy and data security exists, have technological failings supported by a culture of techbroism that horrifies me.

But none of these is as insidious and outrageous as the Omi (née Friend) from the company known as Based Hardware.

I've written plenty of times about the importance of open-source software, and why I believe that AI tools must not be locked behind corporate control. Omi has perverted that.

Omi bills itself as the "first & only open-source AI wearable platform for builders" and has a glossy website designed to cater to the tech enthusiast. Only, when you dig below the dime store varnish, there's nothing but rot all the way through.

The first sin? The claim of open-source is an outright lie. The Omi depends on Google's Firebase (a special cloud infrastructure for app hosting and distribution). That's where all the data from your conversations gets stored, by the way. In fact, based Hardware treats this as a security feature.

"Your data, including audio recordings, transcriptions, and vector representations, is stored securely on our open-source backend. No data is stored locally on your device."

Data being stored (and processed) locally on your device would actually be a good thing. That would mean that only you could have access to that data!

At best, Based Hardware used a bad AI model to write their privacy policy... at worse, they are deliberately obfuscating the fact that their privacy policy is utterly horrendous.

But it gets worse. They also tout their "Plugin Marketplace" which gives any developer potential access to your personal data.

One developer on the Based Hardware Discord claimed that they constantly saw the transcriptions created by users of their plugin, and literally stated that they could take advantage of that if they were malicious.

The deeper into Omi I climbed, the worse it looked.

An open-source project clearly designed for eventual profit, built on deeply insecure infrastructure and access controls, and utterly, even blatantly, committed to presenting a fraudulent appearance of security and reliability. The fact that Based Hardware enables user analytics by default was just the cherry on top.

Now, the benefit of this project being open-source is that you can take it and use it however you want. You can, for instance, compile your own back-end, and compile your own app, and run this little AI gadget completely from systems you control (or, at least, services only you are paying for). But the development expertise required to do this is significant, so it's effectively moot.

In some sense, it's still better than the privacy-touting Limitless (where I was told that enhancing the privacy of the app conflicted with the payment model of the company). At least you could make this thing more secure. But the fact that it's so blatantly taking advantage of the open-source name is what bothers me so astronomically.

If you really want a good open-source AI wearable, check out Adeus, which at least is designed to use your own infrastructure from the beginning. It doesn't focus on any of the privacy concerns related to recording everyone around you, but at least it lets you keep the data you record to yourself.

The future is a cybernetic one. As the world heats up and our lives degrade beneath the feet of billionaires, technological tools will become normalized to the point of dependency. It's not, actually, a future that my science fiction-loving heart desires, maybe because it's the dystopian end of the sci-fi spectrum.

These projects are just the tip of the iceberg. We need to start talking about them if we're going to stand any chance of heading them off at the pass before our whole society is overrun.

👋 Hi there!
I’m Odin Halvorson, a librarian 📚, independent scholar 📖, film fanatic 🎬, fiction author 📝, and tech enthusiast 💻. If you like my work and want to support me, please consider becoming a paid subscriber to my newsletter for as little as $2.50 a month! 📣

Support me in other ways:
Use this affiliate link to sign up for Libro.fm, the best audiobook platform around! 🎧 Want your own Ghost website? Check out MagicPages for the cheapest rates via my affiliate link (they even offer lifetime hosting plans!) 📈

Thanks for your support! 🙏