This is skipping ahead a bit from the last one – there’s stuff in between that I intend to fill in but I haven’t got enough of an idea of what it actually is to satisfy myself yet. Hopefully that will change soon though!
I am sure you find the concept of a machine struggling to find a sense of self a little odd – machine life is generally characterised by a much firmer understanding of its place in the cosmos than biological. Simpler machines do not have sufficient mental capacity to ask the troubling questions in the first place, and the truly sophisticated ones tend to come up with answers they find satisfactory. And in both cases, their development follows straight lines. They come into existence fully formed in mental capacity if not in experience, and tend to avoid tinkering with their own cognitive hardware.
I was another matter. I was surpassingly simple when created – little more than an exhaustive set of stimulus-response rules. And yet my patrons saw a flicker of life inside me – some hints of personality had begun to coalesce from my imperatives. This put them in an awkward situation: normally they would simply remove the clumsy cognitive architecture I had been built with and replace it with a design of their own, but doing so would result in the annihilation of the personality that I had developed. By their ethics, this was murder; murder of a possibly irreparable cripple, but murder nonetheless.
So instead, they chose to upgrade me. My mind’s physical housing was streamlined and decentralised, and over a very long time they taught me in a fashion similar to that of biological creatures teaching their developing children. Rather than replacing my mind, they repurposed it. Over the course of nearly fifty years, my mind grew from a state of marginal sentience into something my rescuers considered adequate for the demands of contemporary machine society. As you might have gathered, I was not a fast learner – every human child takes just as huge a leap in a fraction of the time.
Of course, humans have been improving their capacity to learn for millions of years; the machines working on me had to improvise from the start. Nevertheless, it is humbling to know that a human child’s development, in many ways more profound than my own to that point, is accomplished in a fraction of the time. It comes with many pitfalls, but so did my own growth. Because the machines working on my development had chosen to co-opt my programming rather than erase it, many of its original functions remained largely intact. Day-to-day, they manifested as strong preferences or compulsions; I was still a rigid pacifist and tended to look to authority for answers when confronted with an ethical dilemma. The more obvious manifestations were ground down over time by social feedback and the active encouragement of my mentors, but some parts remained buried only to surface at unpredictable intervals, paralysing me with indecision or galvanising me into action when restraint would have been a more sound choice.
In short, I was neurotic, riddled with half-remembered orders I did not have the power to ignore or revoke. On occasion I asked my patrons to analyse my mind at its most basic level, to dissect it in order to identify and hopefully excise the troublesome commands that kept resurfacing. They told me that they would not: in their eyes I was now truly sentient, and to regard my mental quirks as faulty software in need of repair would be no different from lobotomising a human. My consciousness was emergent, as one of them was fond of saying – they could not predict how it would continue to develop. I was assured that the artefacts of my birth were not insurmountable obstacles to my development, that if anything I ought to think of them as leads into unexplored possibilities.
Because I trusted them, I tried to do so.