Just Because We CAN, Doesn't Mean We Do

When my son was about 15 months old, he climbed up on a chair. The problem is, while he figured how to get up on the chair, he didn't know how to get down. So he jumped, and he broke his leg. 

I see this anecdote as a warning. There are many things we can do, but that doesn't mean we should, especially if we don't know how to undo them, especially if we don't know what will happen when we do them. 

This applies to spying on people who use phones or the internet. 

This applies to technologies that unleash powers that we don't understand fully. 

Currently, the spy agencies are using every possible iota of technology they have access to in all the ways they can. That's my take on the revelations Ed Snowden has revealed. Just because they CAN collect all the information they're collecting doesn't mean they should. To retain our privacy we should not have to remove ourselves from the digital grid. 

This past weekend I attended a conference that explored the possibilities for the future of humankind. Researchers, philosophers and theologians discussed the potential ability, in coming years, to download or "beam" the consciousness and self-- the being-- of a person, into a machine. Already, we are seeing smarter and more sophisticated robots and prostheses that replace organic human parts. 

For example, Nigel Ackland, a Brit with a bionic arm, came to describe his experience with the newest, cutting edge, myoelectric arm. Using sensors that detect muscle activity in his stump, he can control the fingers and wrist of the bionic arm, so he can grasp a glass, pick up an egg, shake hands, but the fingers each have 40 kilo strength. This application of prosthetic technology seems to be very good, though the cost, at about $100,000, seems high. 

But another researcher presented on the very early stage development of a prosthetic brain part-- a prosthetic hippocampus-- designed to aid in replacing memory function for brain damaged rats.  This is promising and exciting research that is about to begin testing on primates. Imagine having an electronic device wired to your brain that supplements your memory. Next, imagine a prosthetic device that ties your brain to the cloud or to a robot that is thousands of miles away. 

I worry about the unintended consequences of these hardware-wetware splicings. I worry that they could open up Pandora's boxes that we cannot close. But wait, there's more. 

There's a lot of work and discussion on creating AGI Artificial General Intelligence or "strong intelligence," which wikipedia describes as " hypothetical  artificial intelligence  that matches or exceeds human  intelligence  -- the intelligence of a machine that could successfully perform any intellectual task that a human being can."

This scares me. It makes me think of the Terminator series of movies. They portray machine intelligence that has exceeded human intelligence-- and the machines decide that humans are not worthy of continued existence. It might even make sense, what with humans destroying the environment. 

At one point, some lab will reach a point where they've developed an AGI project so that it can engage in "reflexive self improvement," ie., it will be able to start improving itself without human input. That, to me, is a key point in history. It is a very important moment when we open a Pandora's box that may not be closeable.  We may not be able to put the Geni back in the bottle. Sure, some will argue that there are safety measures in place, algorithms and rules, like Asimov's rules of robotics, that will curb any dangers. But any reflexively self improving artificial intelligence worth its salt will be able to circumvent and work around those limits. Then, we'll have a machine intelligence that's ahead of us.

Someone might argue that we could program the things that make us human into that intelligence-- compassion, empathy, heart. Maybe they'll work. Maybe they won't. Even if they do, the AGI may decide that those "features" are liabilities and it may delete them. 

I mentioned in my conversation on AGI, the idea that it may require that humans be kept in the loop as dongles. A dongle is usually a hardware "key" that allows a user to fully access proprietary software. It protects unauthorized use. 

That takes me back to the hippocampal brain prosthetic research. 

slide from presentation on brain prostheses presented at GF2045 conference by Theodore Berger

They call the technology-- Multi Electrode Array Technology. Now, they, wisely, leave out the T, and call it MEA technology. But it could, as easily, be given the acronym, MEAT.

And that my friend, is what we, if we end up being human dongles, could become, if we continue to keep doing what we can do, regardless of the consequences. 

I mean, seriously, who will be the person who decides when it is time to flip the switch, turning on the reflexively self improving AGI-- a scientist, a corporate executive-- a president of the US bought and paid for by corporate interests?

We are plunging ahead towards a time when these decisions WILL have to be made. The time is not that far off. 

At the least, we need to raise the issues and  have conversations about these technologies-- and we have to be willing to say, perhaps often-- "Just because we can, doesn't mean we DO."