Source: Elon Musk wants to merge humans with AI. How many brains will be damaged along the way?
The article addresses ethical concerns vs. rate of technological advances, i.e. ‘bandwidth’ picked up by a maximum possible number of ‘neuronal interface connection points’ so far (my words), where Elon Musk’s outfit and solution currently favors bandwidth over the associated risks of a higher invasiveness when implanting the given neuralink device.
The article goes on to discuss primarily medical applications for patients who have lost important brain functions due to impairment and who might regain or gain for the first time functions which the article terms ‘neuroprosthesis’. Per my understanding, Mr. Sigal Samuel seems to say that primarily the replacement of lost or impaired functionality to improve a person’s quality of life appear making such a procedure justifiable in the first place – or let’s say at least easier to “swallow” as far associated other health risks are concerned. Contrary to that, Mr. Musk’s ambitions – and again, according to how I read Mr. Samuel – are of a different nature, where the first, i.e. Musk, appears to fear that we had better caught up in terms of not losing our footing over the rapid advancement of AI in general and how such technologies are at least fit to replace a great number of jobs – or humans altogether in the worst possible outcome. In this open letter that Elon Musk co-signed, the authors raise attention pertaining to these issues:
“AI systems with human-competitive intelligence can pose profound risks to society and humanity” and went on to ask: “Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization?”
Next to the ethical implications I see a different issue altogether and it’s a rather mundane and apparently so far overlooked one per my current understanding of the pros and cons of this technology and humans’ role in it. In order to (hopefully) better get my point across, let’s go back to the wording of the article once more:
Although Musk is not alone in warning about “civilizational risk” posed by AI systems, where he differs from others is in his plan for warding off the risk. The plan is basically: If you can’t beat ’em, join ’em.
Musk foresees a world where AI systems that can communicate information at a trillion bits per second will look down their metaphorical noses at humans, who can only communicate at 39 bits per second. To the AI systems, we’d seem useless. Unless, perhaps, we became just like them.
So again, Musk mentions much smaller bandwidth that human brains are capable of as opposed to what AI can currently do or what it will be capable of in the future. Is it just me or is there a basic lack of understanding how computer technology in general works, at least current classical computing vs. quantum computers, the likes of which we only have prototypes of at this point, maybe campus–wide used ones like the old mainframe back in the days that an entire university’s staff and students shared. What I’m talking about is the way that information is transferred between a computer’s components, in particular RAM, pipelines, parallel processing and such. I take it my readers will be familiar with those basic concepts, so let’s get right to the ‘sore point’ I see: There is so much information that can be handled at a time, say 39 bits per second in a human brain, potentially trillions of bits by a neural network or cloud based infrastructure. Alright, all fair and dandy. The neuralink interface or one of its competitors pick up a certain amount of impulses from a human brain’s synapses and hand over that information to an AI infrastructure, thus in theory augmenting the brain’s capacity of information processing. Is my math really this poor or am I seeing a threshold of 39 bits/s max of input to the AI algorithms? Please work with me for a second here: All the effort that’s currently revolving around developing these devices and make them less invasive whilst more powerful at the same time seek to amp up the capacity of a human brain, correct? (at the very least, that’s the biggest beef one Mr. Musk seems to have with AI at this point) So, if the brain currently can’t “deliver” anything above said 39 bits/s max, then which information is the computing infrastructure going to work with? In order to stress this point further: Say, we had an entire brain “wired up” via intravenously implanted sensoric interfaces that connect to the entire available amount of neural connections in the brain – provided we’ll ever get there, but that’s another discussion for another day. Let’s for a minute pretend we’ve achieved this degree of interconnectivity between a classical human brain and some sort of computer network. Then can this computer network process anything else than the 39 bits/s of “influx” of information picked up by the brain? What I’m asking is: Other than replacing functions that hadn’t been there naturally and thus potentially enriching a person’s perceived quality of life – which is a desirable and IMO ethically justifiable objective – where does the ‘augmentation’ occur? What else can the computer network work with other than the stimuli picked up from and received by that were generated in said person’s brain to begin with? Get my point….? Let’s hopefully help it with an analogy: There’s only so much water flowing down that river. If the river doesn’t carry more water and if no other streams enter into it… how can said river “produce” a higher volume of water being carried along?
Am I hung up on a lack of information and understanding as to the workings of AI in general? Or is it indeed a point never addressed so far in the heat of the argument?
Help me out in the comments, if so inclined. Thank you. And thanks for reading in the first place and indulging my musings, much appreciated.
copyblock.wordpress.com 