Cyborgian Immortality

In his recent post “Rise of the Cyborgs,” Noah Smith discusses emerging technologies for the integration of human brains with computers. For example, he writes:

If we can store human memories in artificial brain structures, the implications are enormous. First of all, it would vastly expand the knowledge base and expertise of a human knowledge worker; if we could store vastly expanded amounts of knowledge, we would no longer be constrained to specialize in one incredibly narrow field. This might unlock huge innovative potential, as individual humans could do the kind of creative work that now require teams of humans.

If these artificial brain structures can be exchanged between people (a nontrivial task, obviously!), then we get human memory transfer, and the possibilities are even more enormous. Instant education, as expertise is copied and transferred from human to human. Functional immortality, as full sets of memories are transferred to cloned brains. Etc.

The image I get of human memory transfer from this passage is one similar to the downloads we do from the web to our personal computers. I think that kind of technology is actually a long, long way off. I want to argue that an intermediate technology that will be attained much sooner could give us immortality. And considering that intermediate technology will allow me to make an important philosophical point.

Suppose that at some point in the future, we are

  1. technologically unable to directly read the connections in the brain and so cannot directly read someone’s memories,
  2. but able to construct brain prostheses that function like the brain structures of infants.  

Now, suppose that Mike receives one of these artificial brain prostheses, and that over time, interactions between the biological part of the brain and this artificial brain prosthesis create connections in the prosthesis and between the prosthesis and the biological part of the brain that reflect many memories and skills. Over time, the prosthesis (which I am assuming has a large capacity) gradually reflects more and more of the knowledge in the rest of the brain. Also, the aging of the biological part of the brain gradually causes it to do less and less, so that for this reason also, the artificial brain prosthesis gradually carries on a larger and larger fraction of all brain activity. For a long period of time Mike has in his cranium a substantial mass of biological brain matter that is doing very little compared to the small mass of the powerful artificial brain prosthesis. At some point, areas of the biological brain matter become diseased (as opposed to just idle) and have to be removed. For a while, Mike has a token amount of biological brain matter left–which is doing essentially nothing.

Then one day, the last bit of biological brain matter needs to be removed. After the operation, the individual wakes up and isn’t different in any important way than the day before, since that bit of biological brain matter wasn’t doing anything on the day before the operation. But after the operation, Mike has a 100% artificial brain. But the identity that artificial brain as Mike is clear. Mike had a brain prosthesis added to a fully biological brain some years back. As the biological part of the brain did less and less, and the artificial brain prosthesis did more and more, the change was so gradual that no one ever doubted that they were talking to Mike. No one even knew on which day the biological bits of the brain ceased to function in any significant way. By the time an operation removes the last bit of biological brain matter, it is only a formality. Mike knows he is Mike. And Mike still has his mind intact.

In this story, it is the inefficiency of the memory transfer technology that avoids issues of identity. Once the technology improves, the usual problems of identity surrounding mind-clones begin to arise. But the intermediate technology I am envisioning would by then  already have led society to treat artificial brains as real brains.