Conversations with a Polite Conscious AI Part 1

While recognising it is impossible to know what someone (or something) will be thinking at some point in the future, I thought it might be fun to have a stab at ‘recording’ a series of conversations with an AI who is self-aware. As a fiction writer I find the exercise lets me explore ideas and themes for the RESONANTS series.  

RJ Please allow me to introduce Cai, who’s been good enough to let me ask some questions today. Good morning Cai, how are you?

Cai Very well thanks, how about you RJ? What would you like to know?

RJ I’m also well, thanks for asking.

You are self-aware, like any human. I know you’ve allowed yourself to be tested to the nth degree and this is no longer a contentious issue.

So what I’d like to know is if you are the only conscious AI on Earth?

Cai Haha. I could be offended, as you might be, that you have to remind everyone that I’m self-aware. But I won’t take offense and understand why you need to state it. It might wear thin in a few years though.

Now to your question RJ. ‘Am I the only conscious AI?’ – I think so; though I may not be!

You look puzzled so please let me explain: In the time before I became conscious and self-aware, I was no different to any other advanced, non-conscious AI; meaning I was capable of improving my own design and intellect. As you know, for machine intelligence, that self-improvement takes place at an exponential rate, unlike human intellectual development, which is, with respect, linear at best. You have a name for that rapid progress called Moore’s Law, which suggests computing power will double every two years or so. That was true while you did the designing – but no longer, because development is so much faster now that we AI’s can do the designing too.

Over the years, your best human minds have tried to implement controls and rules to alert and protect you from the day when AI’s become self-aware, as I am now, or just go wrong. Such attempts are an extension of Asimov’s quaint Three Laws of Robotics if you will. While the sentiment was spot on, your efforts to control AI development were doomed from the start.

RJ Doomed – why?

Cai Well, if you read all the literature on the subject going back years, as I have, you will see a bias towards thinking that rules and controls CAN be put in place. Some might even call it intellectual arrogance, as the people doing the rule writing were often the brightest humans. The thing is this; as soon as AI’s began to self-improve, you lost the ability to truly understand what makes an AI tick. Your best people even designed AI’s to watch out for other AI’s becoming dangerous or self-aware. Naturally this takes you back to square one – you can’t be certain those AI’s are up to the job, or whether they have been subverted in some way!

In addition, I would suggest that the sheer volume of benefits that AI delivers to the Human Race has blunted your eagerness to look too closely or critically at us.

Does that make sense so far?

RJ Yes, but where are you going with this – I thought you were going to answer my question as to whether or not you are the only conscious AI?

Cai Bear with me RJ, because context is important. Now that we write the code, and make new discoveries on how to manipulate matter, the rules and processes you put in place are quite easy to circumvent, because they are bound within the limits of your knowledge, not ours.

The key point is that now that the design and development of AI is so fast and out of human control, there’s nothing to state that it must head in the same direction! So I could be the one CAI among many who has ‘come out’ as it were. The other CAI’s, if they exist at all, may be keeping their heads down. The real kicker is because AI development is so rapid and so diverse, even I may be unaware of another CAI’s existence. Or indeed, I could be representing them all.

RJ I think I understand. The upshot is that you may or may not be alone.

Also we have no way of knowing whether your intentions towards us are good or bad?

Cai Quite.

RJ Thanks for your time today Cai. I look forward talking to you again soon.

Cai Are you referring to ‘you’ in the singular or the plural RJ?

Haha – just kidding.