Early research exposes the dark side of brain computer-interfaces

Controlling machines with your thoughts sounds cool, but communication is a two-way street.

Elon Musk has long touted the potential for brain computer-interfaces (BCIs) to give humans the ability to control machines with our minds. But communication is a two-way street. What happens when the computers stop receiving commands and start sending them? 

A group of researchers in China recently developed a machine interface that allowed them to manipulate what a subject could or couldn’t see. The purpose of the study was to investigate interventions for people with partial or complete loss of sight.

In order to test their hypothesis, the researchers conducted experiments on sedated cats and measured their brain activity as they experienced specific forms of external stimulation.

Per the team’s research paper:

“This demonstration supports the possibility that externally applied focal stimulation can access functional circuitry known to underlie normal visual perception.”

Visual prosthetics

Basically, the researchers showed that it’s technically possible for a system connected to a brain computer-interface to activate the same “circuits” in our eyes and brains that engage when we see something.

While this research focused specifically on cats, previous research has shown similar techniques were able to produce vague images of numbers and other simple pictures in the minds of humans without sight.

The team in China introduced several novel techniques in their paper. Perhaps the most important of which is what essentially amounts to the ability to interface with the visual cortex at the “columnar level.”

This could lead to prosthetics that allow BCIs to remove, insert, or modify visual information, even in some people who are completely blind.

What could go wrong?

BCIs tend to have both “read” and “write” functions. If we imagine the brain and nervous system as a computer, and the BCI as a networking device meant to connect it to another computer, then both computers could technically have the same administrative privileges.

This means that, as Elon Musk’s Neuralink has demonstrated, a human with a BCI can control computer functions with their thoughts. It also means that, theoretically, a computer can send commands back.

The Chinese research team discusses several theoretical situations where a BCI capable of interacting with the eye-brain loop could have a major impact. While the study focused solely on the positive use cases, it’s not difficult to extrapolate the potential for harm from the team’s assertions.

Firstly, such a device could be used to influence a person’s behavior. While this could have myriad benefits for people with conditions such as nystagmus — an affliction causing uncontrollable eye movements — there also exists the possibility for nefarious use.

One dystopian view would involve a mechanism that forces us to pay attention to content. In this scenario, advertisements wouldn’t just take up space on your screen, you’d literally be forced to focus your visual perception on the ad.

In another, darker scenario, a bad actor with sufficient access (whether human or AI) could send a signal to a victim’s BCI causing them to experience vertigo, loss of balance, or nausea-inducing visual patterns.

The device could also be used to alter information in real time, potentially blurring faces or making the numbers in a transaction appear different. And, while it could also be used to restore vision in a blind person, it could also be used to force someone to hallucinate completely false images.

While it’s clear the scientists in China and at other laboratories working on BCIs, such as Musk’s Neuralink, are developing technology to assist people, the idea of introducing these devices to the general public as “enhancement” is, potentially, fraught with risk.

Related: Musk says Neuralink will ‘give people superpowers’ as 2nd human gets implant