Fungi of the rain forest can be nasty parasites. Eager to reproduce, they’ll infect a far larger, more powerful insect, taking control of its brain, and using its strength against it–animating the zombie insect to climb to the far reaches of the rain forest canopy. The insect dies, of course, but the spores are released in the perfect spot, giving the fungus its best chance of living on.
Project Alias is the technological equivalent to parasitic fungus. But instead of latching onto an insect, it latches onto a Google Home or Amazon Alexa device–taking control of their strengths for its own purposes. Project Alias serves as a gatekeeper between you and big corporations. It effectively deafens the home assistant when you don’t want it listening, and brings it to life when you do.
It’s a dramatic metaphor, but an apt one to Tellart designer Bjørn Karmann and Topp designer Tore Knudsen. After all, Google’s and Amazon’s voice assistants are now listening on more than a billion devices worldwide, even sharing them by mistake.
“This [fungus] is a vital part of the rain forest, since whenever a species gets too dominant or powerful it has higher chances of getting infected, thus keeping the diversity in balance,” says Tore Knudsen. “We wanted to take that as an analogy and show how DIY and open source can be used to create ‘viruses’ for big tech companies.”
Project Alias is designed as a completely open-source hardware/software solution for a world where big corporations have the ability to listen to us all the time. The hardware is a plug-powered microphone/speaker unit that can sit on top of your smart speaker of choice. It’s powered by a pretty typical raspberry pie chipset, the tool of choice for homebrew electronics aficionados.
The speaker sounds like a white noise machine to the assistant, covering your speech with an inaudible, omnipresent static. That is, until the software side comes into play. You can train the Alias through local machine learning (no cloud here!) to learn how to wake the assistant to a unique keyword, disabling the static.
The Google Assistant makes you call it “Google.” The Echo makes you call it “Computer,” “Amazon,” or “Alexa.” So instead of talking to something you own, you’re talking to a brand. Alias lets you train it to recognize “Hey Jim” or “Pizza party!” or whatever else you imagine.
“When a family gets a puppy into their home, there is always this name-giving ritual, where the kids get to wish for the name,” says Knudsen. “We don’t see a reason why this should be different with home AIs.”
When you utter your chosen word, it prompts the Alias to whisper, “Hey Google,” to activate the assistant. And then Alias goes quiet, allowing you to communicate with Google or Amazon as you normally would.
The most appealing part of Project Alias is its promise of privacy. Amazon has a relatively poor track record here, storing past conversations in the cloud. Google, too, collects spoken data. Of course they aren’t meant to listen in to your private conversations, but by nature, the devices must always be listening a little to be listening at just the right time–and they can always mishear any word as a wake word. But whether these devices are true privacy invasions or not, frankly, it’s hard to trust big companies with relatively poor privacy track records to always hear only what you want them to hear.
Project Alias offers an independent layer of protection to any privacy-minded person. To be honest, I wish it weren’t just an open source maker project. I wish it were a real product that I could buy right now.
“If somebody would be ready to invest, we would be ready for a collaboration,” says Knudsen. “But initially, we made this project with a goal to encourage people to take action and show how things could be different . . . [to] ask what kind of ‘smart’ we actually want in the future.”