Aardvark DailyNew Zealand's longest-running online daily news and commentary publication, now in its 25th year. The opinion pieces presented here are not purported to be fact but reasonable effort is made to ensure accuracy.Content copyright © 1995 - 2019 to Bruce Simpson (aka Aardvark), the logo was kindly created for Aardvark Daily by the folks at aardvark.co.uk |
Please visit the sponsor! |
Artificial Intelligence is in the news again, as is artificially grown brain tissue.
Despite the fact that several of the world's leading scientific visionaries have expressed strong concerns about the dangers of developing sophisticated AI systems, it appears that we're going full-steam-ahead.
This week IBM demonstrated a neural network built from the company's new "neuromorphic" chips. The "brain" being displayed was built from 48 of these chips and it is claimed to be roughly the equivalent of a rodent's brain.
Each neuromorphic chip (code-named TrueNorth) contains the equivalent of a million neurons and contains around 5.4 billion transistors yet consumes just 70mW of power.
This power consumption is minuscule when compared to a typical microcomputer CPU which, despite having just a quarter the transistors, can require as much as 140 watts to operate.
Using its own "deep learning" algorithms, IBM has been able to teach systems built around the new chips to perform tasks such as facial recognition, word recognition and language comprehension.
This is impressive, albeit slightly scary stuff.
One of the things about neural nets and AI is that increasing the power and capabilities of such a system may simply be a case of scaling the hardware, lighting the blue touch paper and standing well clear.
What happens when they create a synthetic brain using 48,000 of these chips instead of just 48?
In theory, this "brain" would have the potential to develop an intelligence far in excess of the human brain and a cluster of 48K of these chips is quite a viable project.
I don't even want to ponder the potential dangers associated with creating a machine that is hugely smarter than those who created it.
However, shifting from silicon to slime...
Researchers from Ohio State University have succeeded in growing an artificial brain made of real neurons, synapses and stuff.
Fortunately, their goal isn't to create an intelligence but to come up with tissue that can be used to test drugs in a way that doesn't involve using tissue from patients.
So far, their biggest clump of brain tissue is no bigger than a pencil eraser which equates to the brain of a 12-week old human foetus. The researchers are quick to add that this isn't really a small brain, just a clump of brain cells. There is no arterial or venous system as would be found in a real brain but it will offer valuable insight in to the function and response of neural tissue to possible treatment options which can then be applied to people.
What I find most fascinating about this is that despite the fact we may soon be able to build a sentient intelligence, we really have little understanding of the mechanisms behind a consciousness and awareness of self.
The question some will doubtless be asking is "will an artificial intelligence have a soul?"
I guess the answer to this would have to be a resounding "no" -- given that even those who believe in the concept of "souls" would likely believe that such things can only be created by God.
Please visit the sponsor! |
Have your say in the Aardvark Forums.
Beware The Alternative Energy Scammers
The Great "Run Your Car On Water" Scam