But I digress.
One science report claims that thought-reading machines may not be too far off.
A team of world-leading neuroscientists has developed a powerful technique that allows them to look deep inside a person's brain and read their intentions before they act.
The research breaks controversial new ground in scientists' ability to probe people's minds and eavesdrop on their thoughts, and raises serious ethical issues over how brain-reading technology may be used in the future.
My first thought? Puh-leaze.
I'm sure this is either case of wild optimism from the scientists, shoddy reporting from the Guardian (not the Guardian!), or a combination of both. In this case I'm going with a reporter that doesn't quite understand what he was told. If you really read the article, most of the fine grained "reading" seems to be with machines that have been specifically tuned to a specific brain. I doubt they can ever tune a machine with general rules and still be able to pull any meaningful information out a random population.
There is more application as a diagnostic tool. Or as a device to facilitate mind-machine interfaces such as those being developed for artificial limbs. And, as with any new tool, there will be ethical concerns to be dealt with. The sheer invasiveness of such a device would be one such concern. But the subtext of the article appears to be some kind of general thought reading machine that can predict crime is just over the horizon. And I'm at a loss as to how the reporter could reach this conclusion.
Well, not really. I'm sure that part of it he understood, part of it he didn't, and some of it he extrapolated using movie-science (i.e. nonsense).
A study that is able to predict a subject's choice of adding or subtracting two numbers is a far cry from being able to read intentions in general. What this study says is that they can read the intentions of someone in a rigidly defined situation with only two possible outcomes, and I'm guessing only after extensively tuning the machine to each individual's particular brain.
This is way too specific a technique for any kind of general crime-prevention tool. Not only does it need to be tuned, they have to know what exactly they are looking for. And I'm betting while it may never be impossible, it seems it would be so highly improbable that it may as well be considered impossible.
There is no "alphabet of the synapses" where they can observe a person's synaptic activity and look up what it means in some lexicon. Because no one person's brain works exactly like another's. A large part of this is because there's a mini-evolutionary war that occurs when your nervous system is developing. Proto nerves are in competition to become either synapses or glial cells. Due to the trillions of cells involved, no one brain develops in quite the same way as any other.
Just as it is in our macro level ecosystems, so it is in your brain. Nature does not use the ideal model for a situation. It uses the first model that survives. The one that is "good enough". Take the human eye. It most definitely is not the ideal model for an organ that processes the visual spectrum of electromagnetic radiation. Hell, there's a blind spot right in the center of our vision where the optic nerve enters the orb. But it's good enough (and the binocular aspect of our visual field helps compensate). It has contributed to our survival and perpetuation as a species, therefore there is no longer environmental pressure to find a better ocular organ. But you can bet your ass if that little spot caused problems, it would have been bred out of existence long before we developed these big brains of ours.
Now, due to our DNA, every brain has the same essential starting blueprint and overall structure. But there are pressures at work as well and the brain is, essentially, a closed environment. Think of the brain as a ecosystem with trillions of little denizens. It is constantly searching for ways to accomplish tasks "well enough". If a certain pattern of firings or cell configurations produces a desired result that pattern will be used again. And again. Until is becomes more-or-less "hard wired". There is no optimization involved. The result is what matters, not the method used to get that result.
What I'm getting to is that different people end up with different thought patterns and physical structures. To be sure, the meta-functions may be located in the same generalized areas of the brain. Language in Broca's Area, sight in the Visual Cortex. But the specific instantiations and pathways of thought are most likely unique for each person.
And to top it all off, how fine a resolution does this machine have. Say there a million synapses firing in a specific sequence are involved in making the decision to murder someone. Now, what if these same million synapses fire in the same sequence but have been appended by, say, 25 other synapses, and prepended by another 10 or so. But now this represents the intention to buy a cup of coffee. Can this hypothetical machine detect a .0025% delta in brain activity that represents two wildly different intentions?
Of course all this could be moot anyway, depending on whether the brain is deterministic in nature or not. If it is, then perhaps, with the correct tuning for an individual, this could be used as a sort of general "reading" device. But I doubt it could ever be used as a Minority Report style pre-crime detector.
And if the brain operates on a quantum level, then all bets are off. Will the subject be aware of observation or not? Will the observers be aware of what they are observing? Will the quantum states of the observers brains affect the states of the observed? Quantum mechanics tells us yes. If the brain exists as a set of potentials and probabilities, true observation of the brain at a level that the article alludes to will be as impossible as trying to simultaneously measure an electron's position and angular momentum (can't be done).
Basically, is the act of thinking about how I think, somehow, at some level, altering how I think at that particular time?
Shit like this keeps me up at night.
So what do I think (heh) of this? I think the best that will come out of this, in terms of law enforcement, is an extremely accurate lie-detector. Lying is more or less understood in terms of gross brain activity for most people. Pathologic liars and others with sociopathic tendencies will still slip the net. But it will be much better than this bio-feedback voodoo that's in use now.
But reading intentions and specific thoughts? Very, very doubtful.
Feedburner
