Neurophysiologist Martin Zizi claims to have invented a new form of biometric technology that is nearly impossible to spoof, crack or hack.
The system uses the human nervous system as its method of authentication, which transmits signals from the brain to the hand to the sensors on smartphones.
These sensors extract a unique signature from microscopic vibrational patterns in the hand muscle. The user holds the devices for three-to-four seconds before the algorithm authenticates them for access.
Zizi has named the authentication technology Neuroprint, and sells it through the Californian company he founded called Aernedir.
"It is akin to a brain scan from the palm of your hands," he says. "Using kinetic sensor from a normal phone - the accelerometer and the gyroscope - you can measure vibrations. Not motion, but vibrations you don't see that are like shivering in the muscle."
Zizi argues that Neuroprint is more secure than voice, facial, fingerprint or IRIS biometric systems, as all of these can be recorded and copied, and can’t be replaced if they’re hacked or stolen.
Facial recognition can feasibly be spoofed as it’s based on a physiological live signal and not features. Some existing applications including the Windows 10 Hello face authentication system can be spoofed just by using a photo of the user’s face.
Fingerprint biometrics can also be compromised using dummy prints that match the stored templates used by sensors, as Michigan State and New York University demonstrated in a study using computer simulations.
Voice authentication is similarly vulnerable. Scientists from tech giant Baidu claim they can clone someone's voice by training neural networks on just a few seconds of their speech.
The tech behind each of them is becoming more sophisticated, but so are the methods used to break them, such as real-time facial expression transfer that blurs the line between reproduction and reality.
Brain signals create a more complex challenge for hackers.
"Having encryption based on your brain, means that it cannot be decrypted unless you touch your machine, or you have a contraption enabled to scan that and decrypt it," says Zizi, who spent 15 years as a professor of neurophysiology at the Free University of Brussels before he set up Aerendir.
Alternative forms of biometrics can also be hacked from the cloud to which the signals are sent and the information is stored.
Neuroprint doesn’t need any database of digital signatures outside the device. The only location in which the computations are stored is on the device that the hand unlocks, and the signals it measures are unique to each individual, as no two brains are identical.
"That's honestly the only way to give people the control of their privacy, because there is no risk of transmission of information," says Zizi.
The technology is only possible thanks to the incredibly powerful the kinetic sensors in modern smartphones.
They were initially developed for the car industry, where the speed of the sensor that triggers the release of an airbag could make the difference between life and death, and became affordable enough to be used in a smartphone.
Zizi points to his patents as evidence that there's no equivalent system. He’s already had four granted on the technology, and has a portfolio of 17 patents in the pipeline.
The obvious early use case is to unlock phones, but other potential applications range from online banking to unlocking IoT devices.
In future it could even be integrated into the seat of a car, allowing a driver to activate the vehicle using signals transmitted through their gluteus maximus.
“It would just take one minute," explains Zizi, half-jokingly. "If you give me your butt in the seat of the car long enough, I'll turn it into a useful piece of information."