July 13, 2009 – Original Source: h+ Magazine, by Ben Goertzel
Direct brain-computer interfacing (BCI) may sound fanciful, but it’s already a reality — and in coming decades it will almost surely advance dramatically. Neuroscientists are gradually understanding the electrochemical signals by which our brains encode thoughts and feelings; statistical and AI tools are getting better and better at interpreting complex data. The image at right shows one aspect of the state of the art. In an experiment by a group of researchers from the University of Pittsburgh published in a 2008 issue of Nature, a monkey used signals read directly from its motor cortex to control a multiple-jointed gripper with numerous degrees of freedom — causing the gripper to deliver food into its mouth.
Today BCI research is largely driven by the desire to help the handicapped via cochlear implants, prosthetic limbs and the like — but the scope of potential applications is far broader than this laudable but limited market. The entertainment industry is already getting into the picture; there are currently at least two companies (Emotiv Systems and Neural Impulse Activator) marketing BCI devices for video game control.
As BCI technology develops, we can expect it to increasingly serve the function of cognitive enhancement. I’m reasonably good at mental arithmetic and algebra, but I’d take an onboard calculator and computer algebra program any day. A neural interface to Google, Wikipedia and other online resources would be nice, too. And I wouldn’t mind an expanded short-term memory: no more repeating a phone number over and over until I find a place to write it down! Learning a foreign language? Forget the tedium of memorizing vocabulary, verb conjugations and so forth; just plug some flash memory into your cortex and the knowledge is right there. There seems no fundamental reason all this and more can’t occur in the next few decades.
The majority of today’s BCI research involves the connection of various electromechanical devices to the peripheral nervous system, as we’ve seen with cochlear and retinal implants, and artificial arms and legs; or else the readout of a small set of brain-wave-based control signals, as in the Emotiv game controller (covered in h+ issue #1). Only a handful of maverick researchers now explicitly pursue advanced forms of BCI that seek to read more abstract thoughts from the brain. The main bottleneck slowing this research is the lack of adequately accurate devices for measuring and stimulating the brain. In this regard, one critical research direction is the development of safe ways to implant more advanced BCI devices inside the skull. It will probably continue to be easier to read the brain state from within than without, though a breakthrough in “brain imaging from the outside” can’t be ruled out. Scientists are exploring multiple radical brain imaging technologies, including devices involving carbon nanotubes and other nanotech-based materials, which seem to play more nicely with brain cells than conventional materials.
For now, many of our best insights into brain function have come from studies placing electrodes deep inside the brain. Dr. Rodrigo Quiroga and his colleagues have made great progress toward understanding how memories of faces, objects, animals and scenes are stored in sparse neural subnetworks in the region of the brain called the medial temporal lobe. Understanding how the brain stores complex information is step one toward figuring out how to read this information into a computer.
And in time, even more fascinating possibilities may be realized. Consider the “telepathy chip” — a neural implant that allows the wearer to project their thoughts or feelings to others, and receive thoughts or feelings from others. There seems no in-principle reason why this can’t be done, but it raises a huge number of questions philosophically, technically, psychologically and socially. It’s not clear what percentage of a person’s thoughts and feelings would actually be comprehensible to another person — in many cases, you might send your thoughts to someone else only to find them interpreted as 90% gobbledygook mixed up with concepts and images that are recognizable to the receiver. It’s also not too hard to envision some of the social and economic pressures that might arise surrounding telepathy chips. Would you become suspicious if your husband or wife didn’t want to do a telepathy-chip mind-meld after coming home late Friday night? Might you become suspicious of a potential romantic partner who wouldn’t let you peek into his or her mind? What’s she trying to hide? Teams of individuals linked via telepathy chips might achieve far greater efficiency at some sorts of work than any group of detached individuals with similar skill could. Computer programming comes to mind, where the hardest part of the job is often understanding what other people were thinking when they wrote the code that you have to deal with. Social subgroups rejecting telepathy chips could become isolated, backwards communities similar to the Amish today (who, it must be noted, don’t mind their backwardness and isolation at all).
Ultimately, telepathy chips and related BCI devices could lead to the emergence of new forms of intelligence, “mindplexes” composed of independent human minds, yet also possessing a coherent self and consciousness at the higher level of the telepathically-interlinked human group. AI systems could potentially join these mindplexes, reading from telepathy chips and projecting into the user’s minds not just answers to questions, but also original ideas conceived by the AIs that they believe could benefit the humans. Humans who reject telepathic interplay with AIs could be at a significant disadvantage both socially and economically. Nearly any job requiring insight and creativity would benefit from a stream of “push technology” input from a savvy AI. And wouldn’t your date with Jane tonight go better if your natural charming personality were enhanced by a stream of witty anecdotes and sensitive, empathic statements supplied by an AI who has studied Jane’s profile and history in the context of its comprehensive knowledge of human relationships? Potentially all this could lead to the emergence of a global brain spanning human and artificial intelligence.
BCI is early-stage now, and we don’t know where it will lead exactly, but the near-term possibilities are dramatically fascinating and the longer-term ones truly profound.
Ben Goertzel is the CEO of AI companies Novamente and Biomind, a math Ph.D., writer, philosopher, musician, and all-around futurist maniac.