Brain-Machine Interfaces
Tech investor Elon Musk has elaborated on plans for his recently launched Neuralink venture – a company he founded to link the human brain with machine interfaces – which, if successful, raises some fascinating questions about the way we understand and use language.
Elon Musk has in the past made little attempt to disguise his worries over super-intelligent AI , claiming it could either wipe out humanity or relegate humans to mere pets, and now, kicking against the threat of a Terminator- style future, he has detailed plans to make "micron-sized " human-machine interfaces as a step to "counter for Skynet".
With these interfaces, Musk aims to allow humans to communicate their thoughts directly with each other, a process he claims would essentially allow humans to “engage in consensual telepathy”. This may sound very sci-fi but brain interfaces already exist in the medical realm and in an in-depth interview with website Wait But Why Musk outlines some of the thinking behind setting up Neuralink.
"You wouldn't need to verbalize unless you want to add a little flair to the conversation or something … But the conversation would be conceptual interaction on a level that's difficult to conceive of right now," Musk said.
Whether this is possible technically is one question but the concept of this sort of direct brain-to-brain transmission raises some very interesting issues for our very understanding of language.
Science and technology multiply around us. To an increasing extent they dictate the languages in which we speak and think. Either we use those languages, or we remain mute.
J. G. Ballard
First words
The field of direct brain-to-brain transmission has developed rapidly in recent years following the first successful experiments between animals in 2013, as reported in the journal Scientific Report. That experiment, which involved sending signals between a rat’s brain in Brazil and a partner in the US, heralded a new age where the concept of sending brain data between individuals became a concrete reality.Described at the time as an ‘international mind-meld’ the experiment itself is far from what Musk hopes to achieve but points to a future where technology can enable cooperation. In the experiment, the Brazilian rat – the encoder – was trained to press one of two levers to gain a reward, dependant on whether or not an LED in its enclosure was lit. A neural interface recorded the activity in the rat’s motor cortex and then transmitted this to the US rat – the decoder – which was also faced with two levers.
The neural input the US rat received helped it choose which lever to press and, every time it chose the same as the Brazilian rat, gain a reward. Provided they cooperated they could improve their rewards and overall they achieved a 64% parity, far better than even odds and what the researchers describe as “a new central nervous system made of two brains”.
The team conducting the research at Duke University in North Carolina was led by Dr Miguel Nicolelis who believes that this research paved the way to define “an organic computer capable of solving heuristic problems that would be deemed non-computable by a general Turing-machine”
Creating collaboration or multiplying fictions?
As impressive as the hyperbole surrounding mind interfaces is the interesting questions are not so much in the concept of a borg like superbrain but in the ways that humans could ultimately relate to this technology.While rats are able to react to stimuli and adjust their behaviour in order to gain rewards they do not have a developed sense of self and hence their collaboration is fairly straight forward.
Humans on the other hand have developed a highly complex sense of self through centuries of using both spoken and written language. What is remarkable is the potential that neural transmission offers for both language and deception.
It is the ability to manipulate words that first allowed us to cooperate in the plains of Africa and allowed us to generate complicated senses of self and reward/gratification strategies.
Would an implant give us any more idea if someone were lying, for example? Would we not simply learn to game the output for our own purposes?
Our first reaction might be that a link such as this would create - as Elon Musk hopes - an almost telepathic connection, a connected mind that would lead to a utopia of collaborative thought but is this likely? Does the history of natural language evolution suggest this is all that will develop?
What forms of fiction might we see if neural interfaces could send intracortical data between us?
Related articles
A bridge to new language
The rats in the Duke experiment already exhibited some signs of emergent behaviour. Since both rats got a reward each time the decoder chose correctly, the encoder rat started to try and aid its partner in the US by adjusting its movements to create a clearer signal. Over the course of the experiment the Brazilian rat refined its movements making clearer, smoother presses on the lever. In this case, the system was set up to favour collaboration but what would the result be if only one rat could receive a reward each time? Would the Brazilian rat try to obfuscate its mental signal?When it comes to human social interactions there are of course a far wider range of options than simply ‘left’ or ‘right’ lever. Some people will blurt out whatever is in their head while others show icy restraint, some people speak plainly while others always rely on irony, some people invariably tell the truth while others lie incessantly.
"It begs the question - what forms of language will this lead us to?"
Would intracortical microstimulation make these variations less pronounced or more? Would an additional sensory input lead to fewer lies or more? Before the first written language, human cooperation was limited but so too was organised religion or nationwide warfare. Certainly written language has done little to reduce the amount of fiction in the world.
It begs the question - what forms of language will this lead us to?
Greeks Bearing Gifts
These questions all come before we even consider the software and hardware architectures used to transmit any ‘thoughts’. The researchers in the Duke University experiment used software to try and ‘clean’ the signal.With a choice of just two levers the desired result was fairly obvious so they were able to boost the signal-to-noise ratio but what implications are there for the interference of software in transmission when the moral or ethical outcome is less clear cut?
Will Elon Musk develop algorithms to ‘clean’ the transmission between two humans? Should he? What level of trust would we have with a sensory input from such a transmission if we felt that it had been manipulated or ‘cleaned’ by software?
Whatever Elon Musk’s intentions it is undoubtedly many years before anything approaching a neural bridge can be developed for humans but it seems certain that the language with which we, as a society, construct our thoughts in the present day is never more important – as this will be the language upon which these first prototypes will be built.
By  Lochlan Bloom | Embed |
Lochlan Bloom is a British novelist, screenwriter and short story writer. He is the author of the novel the The Wave  as well as the novella Trade and The Open Cage. He has written for BBC Radio, Litro Magazine, Porcelain Film, IronBox Films, EIU, H+ Magazine and Calliope, the official publication of the Writers’ Special Interest Group (SIG) of American Mensa, amongst others. | |
0 comments:
Post a Comment