Introduction: HacKIT: a Civic Privacy Hard(wear) Kit for Hacking Alexa, Google, and Siri
Tired of your "smart" devices eavesdropping on you? Then this surveillance-hacking toolkit is for you!
HacKIT is a low to high tech civic privacy hard(wear) kit for the redesigning, hacking, and reclaiming of the Amazon Echo, Google Home, and Apple Siri. The voice devices are fitted with 3D printed “wearables” and a sound-generating circuit that obfuscate and confuse speech recognition algorithms.
It uses speculative design as a form of civic resistance to empower makers around the world to subvert the hegemony of surveillance capitalism. It puts an exclamation mark on surveillance and exposes the extent to which we've been outsmarted by our "smart" devices.
My hope is that this will be a toolkit to critically-make/create/invent humanity-centered technologies and futures. Have fun!
Download CAD models for the 3D printed "wearables" and audio samples here.
Step 1: (Method #1) Tactile Hack
HacKIT comes with 3 methods of hacking to cater to a wide range of makers. Tactile hack is a low to no-tech method of muting and silencing Alexa.
Materials: Soft-molding clay, rubber foam, copper fabric
How to use: As seen in the images above, use the materials to cover microphones of your voice devices
Outcomes: Voice recording is muffled and muted
Step 2: (Method #2) Algorithmic Hack
This hack aims to feed the speech recognition algorithms of Alexa, Google, and Siri with fake data to disrupt their abilities to build a targeted user profile. Audio loops play on repeat triggering false data recognition. Some audio samples include white noise xx
How to use:
Step 1: Download audio files and transfer to soundboard
Step 2: Download CAD files and 3D-print "wearables"
Step 3: Solder loudspeakers, battery charging port, and switch to soundboard
Step 4: Assemble circuit with 3D wearables and you're done!
Outcomes: Speech recognition algorithms can't accurately develop your user profile, protecting your identity and privacy
Step 3: (Method #3) Obfuscation Hack
Method #3 provides makers with more flexibility in the specificity of their hack. While the Algorithmic Hack plays audible sound loops, the obfuscation hack allows the use of ultrasonic frequencies above the human hearing range to obfuscate the audio recordings of Alexa, Google, and Siri. For that, I've included a PCB I designed and built. Makers can also build off Bjorn's work to customize wake-words and reduce false triggers. Project Alias also allows users deactivate the white noise with the wake-words.
Materials: PCB for ultrasonic frequencies, ATtiny45, 2 Class-D Audio Amplifier, Raspberry Pi (optional), 2 miniature loudspeakers, lipo battery and charger, switch, 3D printed "wearables", Arduino code to program ATtiny45
How to use (ultrasonic frequency):
Step 2: Download Arduino code and program ATtiny45
Step 3: Download CAD files and 3D-print "wearables"
Step 4: Solder loudspeakers, battery charging port, and switch to PCB
Step 5: Assemble circuit with 3D wearables and you're done!
How to use (Project Alias): Refer to Bjorn's documentation here
Outcomes: Microphones of the Amazon Echo and Google Home are obfuscated with ultrasonic / white noise frequencies which prevents any unwarranted voice recording when a user isn't actively using their device. It provides users with both the practical functionality of a voice assistant while also protecting user privacy!
Step 4: Hack Away!
This project is but one of many instances to resist and subvert surveillance in an age of surveillance capitalism. Designers, makers, and technologists have an ethical role to play in dismantling and exposing the black box of surveillance. My hope is that future hackers will iterate, add, edit, and build on top of this work.
“Can speculative design take on a social and possibly political role, combining the poetic, critical, and progressive by applying excessively imaginative thinking to seriously large scale issues?” - Speculative Everything: Design, Fiction, and Social Dreaming (Dunne and Raby, 2013)