The future of frictionless interaction with technology.

23/05/2017
Marin
Cicada Innovations’ resident and founder of BioFoundry, Meow-Ludo Meow-Meow, weighs in on the future of frictionless interaction with technology.
Article by Paul Biegler. Published on 21 May 2017 in the Sydney Morning Herald.

 

Circumstances are making it hard to see what’s in the little plastic cylinder molecular biologist Meow-Ludo Meow-Meow is waving in his hand, with – somewhat ironically given his moniker – the air of an exuberant puppy.

For starters we’re on Skype and his room in Sydney’s inner west is dimly lit. Then there’s his vaping smoke wafting over everything.

“That,” says Meow-Meow, with a triumphant rattle of the tube as the mist clears, “is an Opal card for going on public transport.”

“But I’ve got to get it implanted first.”

Meow-Meow, Bachelor of Molecular Genetics, former Science Party candidate and co-founder of Sydney’s Biofoundry is a grinder, someone embedded deep in the Mad Max-esque world of biohacking for whom putting Sydney’s equivalent of the Myki under your skin is just one more step on the road to transhumanism.

“The rules say it remains the property of CityRail,” says Meow, fingering the chip in a bottle. “I’m worried they could confiscate it now. But if it’s under my skin, good f—ing luck to them.”

Meow-Meow already has a chip in his thumb that can open a door and tell his smart phone it’s him, but there’s a new venture in the world of implants that makes Meow’s invisible hardware seem tame, and it’s brought to you by that entrepreneur with a habit of making mad science real.

In March the Wall Street Journal breathlessly announced that Elon Musk, billionaire purveyor of Tesla cars and Powerwall batteries, and whose SpaceX outfit recently brought Mars colonisation closer with the first ever launch of a “re-usable” rocket, was hiring for a new company.

Elon MuskElon Musk predicts AI will surpass human intelligence “by a lot” – but a brain computer interface could help us keep up. Photo: AP
Neuralink will build a brain computer interface (BCI) or “neural lace” that will eventually “upload thoughts” to the internet but, along the way, deftly heal those with epilepsy, Parkinson’s and depression, and restore function to people with stroke and brain injury for good measure.

Neural lace, says Musk, will add a digital layer to the brain that can wirelessly beam data from our noggins to connected devices and the cloud.

Anyone else would be accused of overreach, but Musk’s track record commands a certain respect, and the job descriptions on his Neuralink website are sufficiently techno-opaque to suggest a very advanced team.

Elon Musk Elon Musk’s Neuralink aims to build a brain computer interface that can “upload thoughts” to the internet. Photo: Susana Gonzalez
So it’s not unicorn-chasing to wonder why Musk thinks it’s a good idea to re-engineer our grey matter and, for that matter, just how the finished product will work.

The back story is that Musk sides with the transhumanist sentiment that, to steal futurist Ray Kurzweil’s book title, “the singularity is near”.

Inventor and futurist Ray Kurzweil pictured in his Boston office. for Insight att: Duska Sulicich...to go with Nick Miller story....photo by Trevor Collens 25 Oct 2011Futurist Ray Kurzweil warns of the singularity – the point at which artificial intelligence starts teaching itself. Photo: Trevor Collens
That’s the point where AI starts teaching itself, prompting an exponential leap in machine smarts that will make humans look, in Meow’s words, “like the Amish”, and in Musk’s parlance, like “house cats”.

Actually, it’s worse than that.

Here’s what Musk said in Tim Urban’s epic April blog post on Neuralink: “A house cat’s a good outcome, by the way.

“AI is obviously going to surpass human intelligence by a lot …There’s some risk at that point that … the AI goes rogue.”

Neural lace will, according to Musk, head off AI armageddon by plugging us into evolving machine intelligence and keeping us ahead of the game.

Of course, as Maureen Dowd noted recently in Vanity Fair, Musk’s dark prophecy could well be tinged with the entrepreneur’s supreme marketing flair – hi-tech cross-branding with a thought uploader won’t hurt Tesla’s image one bit.

Even sceptics, however, can’t ignore feats such as DeepMind’s (the Google AI system) 2016 trouncing of world champion Lee Sedol in AlphaGo, the board game many thought was an impregnable bastion of human intuition.

But if Musk is really concerned about an AI uprising, why start out by curing a bunch of sundry, albeit devastating illnesses, rather than crack on with supersizing human brain power?

The reasons are deeply practical. To implant an experimental device you need approval from the US Food and Drug Administration, and that’s only going to happen if you’re treating a serious illness.

Moreover, as Gerwin Schalk, Deputy Director of the National Centre for Adaptive Neurotechnologies in Albany, New York, told me, each experiment with a brain implant for human use costs tens of millions. And Musk is, after all, a businessman.

“At some point in the future, brain implants will fundamentally change humanity. But at present, the sexy stuff is not it,” says Schalk, referring to the scant foreseeable returns on implantable superintelligence.

“Even Elon is not going to have the amount of resources that it would take to basically ignore economic realities.”

Gerwin SchalkGerwin Schalk of the National Centre for Adaptive Neurotechnologies says treating disease is the first goal. Photo: Mike Wren
Those realities lie firmly within the perimeter of disease, believes Schalk, who sees stroke rehabilitation as a frontrunner application for any commercially viable neural lace.

So what could Musk’s lace possibly look like?

The basic brief for any BCI is to record the electrical activity of brain cells and sometimes squirt back a little voltage to stimulate those same cells.

EEG records brain waves from outside the skull, but high fidelity means getting up close, generally by draping an array of 100-200 electrodes over the dura, the brain’s hard outer layer, or plunging electrodes directly into the squishy stuff.

These deeper brain dives have, already, produced some astounding results.

Deep brain stimulation has helped people with uncontrollable tremor from Parkinson’s hold a cup of coffee. In epilepsy, the NeuroPacedevice senses an impending fit and heads it off with a judicious surge of current. A DARPA program culminated in a paralysed woman using her thoughts to control a robotic arm and gently shake hands with an incredulous interviewer.

And a February article reports that a woman paralysed with Lou Gehrig’s disease used the BrainGate device to type at a record eight words per minute. With her thoughts.

There is, however, a big problem with sticking things in the brain.

“You have an electrode that’s hard and artificial, it’s made of metal and it’s spiky. And on the other side is the brain that’s soft, chemical, electrical and pliable. Those two things don’t work together all that great,” says Schalk, with neuroscientist deadpan.

Brains get traumatised, scar tissue forms around electrodes, the signal degrades, the body rejects the device, and so on.

Which is why the work of a young Melbourne neurologist named Tom Oxley has the tech world slightly aquiver.

Last year Oxley published a groundbreaking proof of concept paper in Nature Biotechnology for a BCI that could record hi-definition brain activity without having to open the skull.

His technique was to wend a device, the “Stentrode,” up the jugular of anaesthetised sheep and plonk it in one of the large veins that lies against the brain, where it sat happily for up to 190 days as the sheep, now awake, went about their business.

In April, Oxley’s company Synchron nabbed $US10 million in funding for a first-in-human trial slated to start in Melbourne next year, with the ultimate aim of having paralysed people walk again by controlling an exoskeleton with their thoughts.

Oxley’s short term goals are more modest, however, and focus on software to decode thoughts and let people drive a wheelchair or control heating, cooling and appliances in the home.

Oxley’s take is decidedly more pragmatic than Musk’s “thought upload” mantra.

“You can get a bit philosophical about it but the thoughts in themselves are only useful in so much as they are recognised by the computer reading the program, at which point an output is triggered. Think about it as uploading a real time command control system.”

Oxley’s vision of “brain as remote control” means we need to be able to switch bits of the Jello on and off. And there is one region particularly well suited to the task.

“You can activate different parts of your motor cortex by thinking about moving particular parts of your body in ways you have grown up all your life doing,” says Oxley.

Then the trick is programming the computer to see that brain activity as a signal to do another task.

For a person with tetraplegia, bumping up the central heating could, theoretically, be just a matter of thinking about kicking a footy, and take your pick of other imagined movements to switch on the microwave

Musk has hinted that his device could also use a vascular route to reach the brain, and so the Stentrode may well give us a peek into what a neural lace could look like.

But the Stentrode’s current command and control interface seems clunky compared to the specs needed for what Musk has in mind.

The tech magnate thinks humans have an “output” problem – our senses import swags of data (think how much you take in with a single visual sweep of a bustling city street) but we transmit at the slovenly speed of two thumbs tapping (try texting everything you saw to a friend).

Musk’s vision calls for outputting data to devices at the speed of thought, and that’s where Schalk’s research may well be on the money.

Schalk monitors people with epilepsy who have had brain electrodes inserted to detect where seizures start and to guide neurosurgical treatment.

He’s developed a computer algorithm that can work out what people are reading aloud, JFK’s inauguration speech for example, just by analysing their brain waves. And in a 2016 article, he used a similar technique to discern words, including “cowboys”, “swimming”, and “python”, that people were merely imagining saying.

The mind-boggling potential is that a computer could read your thoughts just by analysing your brainwaves, a game-changer for people “locked in” by paralysis so severe that thoughts are their only option for controlling the environment.

But Schalk says there’s a conceptual road block to his “brain to text” algorithm linking thoughts and machines any time soon.

“You are not thinking in mental sentences,” he says.

“Language isn’t captured as a series of characters you can somehow find in the brain. It’s a combination of many experiences and many emotions that together realise language in all its complexity.”

Take “red rose” says Schalk. It could be a flower, but it could also be the first movie you saw with your girlfriend (if you like Hindi thrillers) or the name of a forest you once visited that evokes tinkling streams and the chirp of birds.

And if Urban’s blog post is a guide, Musk really is aiming at getting this kind of “all over” thought up into the cloud.

Imagine, writes Urban, you’re on a hike and want to share it with your hubby.

“No problem — just think out to him to request a brain connection … now his vision is filled with exactly what your eyes see, as if he’s there. He asks for the other senses to get the full picture, so you connect those too and now he hears the waterfall in the distance and feels the breeze.”

Back in the present, however, there is an elephant in the room full of would-be transhumans.

Musk himself says we’re “already cyborgs” – our smartphones do much of what is promised by his “digital extra layer” – so why not just share that hike by uploading a video to Instagram? What’s the big deal about bringing the hardware on board?

Meow has struggled to come up with good reasons why he couldn’t just have his chip on a ring.

“At a practical level, having it under your skin means it’s waterproof. You don’t want to wear jewellery at a public pool or a beach,” says Meow, who also envisages one day bringing up his daily diary remotely on a “smart mirror”, while taking a shower.

But the most obvious benefits are, again, for the disabled.

Oxley points out that tasks most of us take for granted, such as moving a computer cursor, may use up a paralysed person’s last remaining function, if they have to manipulate a joystick with their mouth, for example.

“There is a vast region in the cortex that is still functioning tissue, but it is not being used,” says Oxley.

The neurologist sends me a video from a 2015 article in Scienceshowing a man who has repurposed some of that cortex to move a cursor with his thoughts, freeing up his mouth to chat with the woman next to him.

“That is world changing for a patient with severe paralysis,” says Oxley.

But brain implants are also lifting the lid on an ethical Pandora’s Box.

“Some patients seem to experience self-estrangement. They don’t recognise themselves after being implanted,” says Frederic Gilbert, a philosopher at the University of Washington.

Gilbert interviewed a woman with epilepsy whose symptoms improved with an implant, but who felt worse. Having a 24/7 brain box monitor gave her the miserable feeling of being a “full time” epileptic, something she was previously only reminded of intermittently, after a seizure.

Others were weirdly drawn to think they had superpowers.The modern day bionic cyborg?

Gilbert recounts a woman in her 50s who had improved with an implant for Parkinson’s, but one day decided she’d move the pool table all by herself, injured her back, and ended up in a wheelchair for two months.

And, echoing the insubordinate computer HAL in 2001, Gilbert also worries about when “the device will be allowed to kick you out of the decision loop”. Imagine, he says, a military pilot who experiences depressurisation but thinks he is still competent to control the weapon system. The BCI begs to differ and overrides to fire the missiles.

But perhaps the biggest hurdle for Musk’s vision, articulated in a recent MIT Technology Review piece, is more prosaic: who, after all, is going to get brain surgery to send an email?

“I would. F—ing oath I would,” chips in Meow, a self-described neophile.

“I want this technology to succeed, so I’m happy to be a beta tester. As long as I’m happy with the safety.”

Why? In the short term Meow is enamoured with “frictionless interaction with technology”. But longer term, he’s with Musk.

“We have to convince robots and AIs that we are worthy of joining the singularity with them. If we don’t join with machines now they’re going to overtake us and we’re going to be like a cockroach to them … we’re doomed.”

Twitter Auto Publish Powered By : XYZScripts.com