Mind-reading machine
This doesn’t sound like a good idea to me:
…[A] new artificial intelligence research project coming out of Japan…can analyze a person’s brain scans and provide a written description of what they have been looking at.
To generate its captions, the artificial intelligence is given an fMRI brain scan image, taken while a person is looking at a picture. It then generates a written description of what they think the person was viewing. An illustration of the level of complexity it can offer is: “A dog is sitting on the floor in front of an open door” or “a group of people standing on the beach.” Both of those turn out to be absolutely accurate.
“We aim to understand how the brain represents information about the real world,” Ichiro Kobayashi, one of the researchers from Japan’s Ochanomizu University, told Digital Trends. “Toward such a goal, we demonstrated that our algorithm can model and read out perceptual contents in the form of sentences from human brain activity. To do this, we modified an existing network model that could generate sentences from images using a deep neural network, a model of visual system, followed by an RNN (recurrent neural network), a model that can generate sentences. Specifically, using our dataset of movies and movie-evoked brain activity, we trained a new model that could infer activation patterns of DNN from brain activity.”
Nothing to worry about, he says, because it’s a long way from having any real-world applications.
Ah yes, that’s what they always say.
It all reminds me of a wonderful Ursula LeGuin story entitled “The Diary of the Rose,” which I’ve previously written about in a somewhat different context:
…[The story] appeared in [LeGuin’s] book The Compass Rose, one of my favorite collections.
This particular story is about a futuristic society in which a machine that images people’s minds and thoughts has been invented, and is being used for political indoctrination and control. People whose minds deviate too greatly from the PC norm are sent for a type of electroconvulsive therapy that destroys their memories and personality.
Here’s an excerpt from the story, which takes the form of a diary by the worker who operates the machine and writes reports on the brain readings:
Half-hour scope session with Ana J. at 8:00…
It is amazing how banal most people’s minds are. Of course the poor woman is in severe depression. Input in the Con dimension was foggy and incoherent, and the Uncon dimension was deeply open, but obscure. But the things that came out of the obscurity were so trivial! A pair of old shoes, and the word “geography”! And the shoes were dim, a mere schema of a pair-of-shoes, maybe a mans’ maybe a woman’s, maybe dark blue maybe brown. Although definitely a visual type, she does not see anything clearly. Not many people do. It is depressing. When I was a student in the first year I used to think how wonderful other people’s minds would be, how wonderful it was going to be to share in all the different worlds, the different colors of their passions and ideas. How naive I was!
I highly recommend the story, and the whole collection actually.
All the people that want to make fun about MK Ultra, better start now before you become the minority here and elsewhere.
Same with those who rejected out of hand the proof and idea that HRC was rigging the election 2016.
Nothing to worry about, he says, because it’s a long way from having any real-world applications.
The military is usually at least 10 years ahead of what is released to the civilians. US military have shut down University experiments because they claim “it’s classified”.
In Japan, this system is ideally thought to be used in the manipulating of machines (mecha) using indirect human thought control.
The US has already made this a reality with Drone Warfare. AI Drones are too risky, but human controlled drones are a nice fail safe, but require manual controls.
The technology described is a little weird in that it combines two very complex technologies; brain activity mapping to imagery determination and imagery to English language conversion.
I could be wrong, but the first technology seems like a much more difficult task than say lie detection. What if they could make a virtually perfect lie detector. Imagine the uses and abuses!
_____
I once heard a science colloquium speech by two IBM researchers about mapping brain activity without MRI or fMRI. They began with SQuID magnetometer technology. SQuID = Superconducting Quantum Interference Device, the worlds most sensitive magnetic field or electric current detector. Then they combined hundreds of tiny SQuIDs into a super directional antenna for magnetic fields.
There were more hurdles, but they got it to work; though I think the fMRI has better spatial resolution. The plus was that the device was completely external to the body or head, instead of having to crawl inside a giant MRI superconducting magnet.
This doesn’t sound like a good idea to me:
neo: Nor me. I won’t even do a Facebook account, which is miles less intrusive and problematic.
Re: Ursula Le Guin. I just watched “Lathe of Heaven” a few weeks ago. Still holds up. Still brilliant in its low-budget way.
As I recall, Le Guin spoke of “Lathe” as an homage to Philip K. Dick. She was right about that too.
Sadly, Dick was so beaten-down and paranoid (partly all the speed he had been taking) he became estranged from Le Guin.
Le Guin’s “The Ones Who Walk Away From Omelas” was a short story which changed my life. I can still remember staring out a cheap apartment window after I read “Omelas.” It was a Rilke “You must change your life” moment.
She died last January at the age of 88. RIP.
I would like to see this technology replicated widely by teams from multiple countries before I start crediting it too much. It smacks a little of Clever Hans to me.
DNA, Protein research, and mind wave patterns was indecipherable data to computer models until the existence of quantum qubits or quantum computers. Up until now, people have been hacking it with massive networked humans to do the analysis. That has already bottlenecked to a certain extent, as the number of high IQ humans with the capability and motivation to do such things is not that high. It is also extremely boring except for the OCD people.
Major breakthroughs won’t happen even if sensor technology gets sensitive enough to obtain the data, because there won’t be enough quality/quantity analysis of it. Same issue with intel agencies sourcing their intel from FB. It is not that there is not enough data, but that there is too much and not enough analysis humans/software to crunch the data into usable form.
That is why the key import line in the reading is not mind reading but “AI”. It’s the AI that is the crucial solution keystone.