The Future of Computing Part 7: Conscious Machines?

Conscious, Emotional machines, will we ever see them? How far can technology go and can technology be applied to us? In this final part I wonder into the realm of Science Fiction. Then, to conclude the series I come back down to Earth to speculate on the features we’ll see in any radical new platform that appears. Update: Never let it be said I ignore my errors, in the interests of clarity and with apologies to Extreme Programmers I have revised Part 1.

2001, Bladerunner, The Matrix, Alien and Star Trek (TNG) [1] all have machines which at least appear to be conscious and emotional. The idea is so ingrained in us that we expect an Artificial Intelligence to be a carbon copy of a human being, will it be? Will AI computers be conscious? Will they have emotions?

One of the biggest questions about artificial intelligence concerns the matter of consciousness, or “self-awareness”. The entities I spoke of in part 6 may have many of the attributes of living beings but they are are still “machines”, they are not conscious and they do not have emotions.

Will machines become conscious?

There is no answer to this question since we do not at this point even understand our own consciousness or how it works. How can we add consciousness to a machine when we don’t understand it ourselves? How do you program self-awareness? Can it even be programmed?

On the other hand, perhaps programming wont be necessary…

It has been theorised that consciousness is an “emergent behaviour”, a behaviour which appears out of the complexities of an existing system. Could consciousness be a side effect of intelligence and if so will it’s development in our entities be inevitable?

A biological brain is very different from a silicon one. Physicist Roger Penrose [2] has written that quantum effects are present in the human brain and these are important as they could be responsible for non-algorithmic functions. These effects will not show up in a digital system. An analogue system can be simulated but Simulating quantum effects is likely to be impossible and if they are important just faking them may not work. Are these quantum effects important for consciousness to be present? If so purely digital entities may never become conscious.

If a machine does become conscious it becomes something else entirely, something you can have a conversation with, something which may even have a will. Something which may no longer be interested in it’s masters bidding. Something which could turn against us and attack.

Conscious Problems

A conscious system brings up all sorts of problems, will it be emotionally driven in the way humans are? Will it even have emotions? If so how will it react to them?

This is the question behind Bladerunner where the “Replicants” knowing they have a limited life-span go back to their creator in an effort to try and stop their demise. They do not have an entire life to get used to emotions so are highly immature when it comes to dealing with them and consequently tend to leave a trail of bodies in their wake.

An interesting aspect of the film is that the Replicants are considered as non-human and have no rights. Some people have already demanded human-like rights for the more intelligent animals so what will be the situation for a conscious entity? There is no problem if they are not conscious, but if they are (or at least appear to be) should we award them human-like rights? What would have been master director Stanley Kubrick’s last film, “AI Artificial Intelligence” deals with the question of a conscious artificial entity who has very human emotions but no human rights.

“His love is real, but he is not”[3]

Emotional Machines

One thing I think fiction gets wrong is the idea of emotions in artificial entities. It’s by no means clear that they will have emotions and even if they do they are likely to be very different from ours. We are biological entities and there are many factors which effect us and our emotions.

Being chemical based means chemicals effect us. Food, drink, the temperature, the humidity and all manner of other things can change our mood and in extreme circumstances can sometimes radically change our behaviour. For the most part most of us are completely unaware of the effects that many day to day chemicals can have on us. If you are a big coffee drinker try going without for a day and notice how agitated you become, on the other hand notice how chocolate can make you happy (chocolate was once considered a dangerous drug). Foods are full of chemicals and they can have unexpected results, indeed some migraine sufferers have unexpectedly traced their suffering to specific foods.

We also have “programming” of sorts except we don’t usually refer to it in such a manner. Curiosity, reproduction and survival are all part of our programming. These effect our emotions in a big way and pretty much dictate large parts of our lives. If you are interested in a pretty girl is this because you took a conscious decision to be interested or is it because you are simply following your subconscious programming to reproduce? Maybe we’re not so different from machines after all.

“I met someone who looks a lot like you, She does the things you do, but she is an IBM”[4]

The human entity is a complex system. The artificial entity is a completely different form of complex system which will be effected by different things. I do not believe they will act like us and I’m not convinced they will have emotions, if they do I am convinced they will experience them in a very different way.

They may be conscious but I don’t know if anyone can really say one way or the other right now.
I don’t think we’ll have too much to worry about from emotional computers though, I expect they’ll be about as emotional as Star Trek’s Mr Spock. If they have emotions it’s because they’ll fake them.

Unreal Capabilities

The Science Fiction AI entity ends up very similar to us but in reality they will be able to act in ways we can’t do and wont be expecting, they are not human and I don’t expect them to act anything like a human unless they have been specifically designed to.

I talk of an entity as if it is one thing but it will be able to hold conversations with different people simultaneously, a single consciousness may not be able to do this but our entity can split itself apart. It will be able copy itself, it may take over mechanical systems and take on a physical body, it may do the same in the virtual space and take on a virtual body.

They will also do things we can’t even conceive of doing ourselves such as sending memories directly to one another. We can co-operate with one another, these may be able to not only co-operate but actually merge into one entity, then de-merge with all the knowledge of each other. This is just the beginning, they will have capabilities we can’t even imagine. Who knows, Perhaps one day they be able to merge with us.

Cybermen

In part 6 I mentioned cyborgs which were a combination of machine intelligence and a human body. What if it were to do it the other way around and use a human intelligence instead?

Interfacing directly with the human brain is a long term goal and it’ll be worked on for many years so how this will operate is more than a little open to question, but it will give some intriguing possibilities. If our interface to the world goes via a computer we to will be able to take on other bodies (providing the necessary nutrients and oxygen are still supplied to our brain).

The idea that some already have of recording a persons entire lifetime could be turned on it’s head and allow you to experience someone else’s life, rather than watching someone getting shot on TV you could join in and experience it yourself, with the pain turned down of course. Of course sex change operations will be rather simpler, you can become whoever (or whatever!) you want. If you don’t like it just change back. This is a vast oversimplification of course but I am talking SciFi technology here…

If we are injured we just switch off pain and perhaps send a bunch of nano-technology machines to go fix the damage. If we could do the same to the brain (which has no self repair mechanism) we could significantly extend our lives, perhaps even permanently.

Of course the idea of brains without bodies isn’t new, H. G. Wells had Martian brains with mechanical bodies fighting Humans in “The War of the Worlds” way back in 1898?. [5]

Putting our brains into a different body is one thing, but can we fit a human and computer intelligence together? I’ve no idea how we’d communicate but it’d be very useful to have an intelligent computer with you especially which all it’s knowledge and ability to upload more.
Of course in any mixed intelligence we’d need to ensure that the human was the boss.

If the entity is using evolutionary programming perhaps it can find a way to apply it to us, to evolve us. If a way is found to replace biological neurones with artificial ones who knows what’ll be possible. Just make sure there’s no power cut! A human brain made out of artificial electronic neurones will have some of the same advantages of an artificial entity, no need to sleep. It will also have the disadvantages, a completely different experience of emotions or worse, no emotions.

I for one don’t believe we will ever be able to upload ourselves to a computer as is sometimes suggested in Science Fiction so I don’t think the above is possible, but that said I don’t think anyone can really give a meaningful prediction on that matter today.

Artificial Questions

AI poses many more questions than we can answer today. Some answers will be answered when we have AI, others we will be forced to answer ourselves when it has it’s inevitably huge effect on our society.

The future could be a wonderful place and AI could be a very useful part of it but we need to get it right. Release a learning machine too early and it could have very bad consequences for all of humanity. We need to define it’s role and the rules it needs to follow, we also need to extrapolate to see what happens when we push those rules to their limits and when they collide, otherwise we could get unexpected results. Whatever way you look at it the coming of AI is going to change the face of human society.

The normal path of evolution is for superior species to survive the lesser ones. Today we are the superior species, will AI change that, or will we fuse with the machines and ourselves take on a different form of being?

I am of course missing a lot of detail here and am making a lot of assumptions. There are many questions to be asked and technologies to be developed so these possibilities may take centuries to be realised if they are realised at all.

Predicting the future is difficult enough but I have gone into the unknown here and this is a lot more difficult than simple extrapolation. While I may sound like a Science Fiction writer all of these predictions are at least partially based on today’s technologies being combined and extrapolated over a long time period. I’m describing a path, but I don’t know if it’s the one that will walk or even if it can be taken.

Perhaps you can give very good reasons why these ideas cannot possibly work, maybe they break the rules. Those reasons may be valid today, but tomorrow? Science has a long history of rewriting the rules.

I will make one solid prediction though:

You’ll know they’ve got Artificial Intelligence right when your robotic girlfriend says No.

Now, Back to Planet Earth for the Conclusion and a New Platform…

In the future computers as we know them today will become playthings for geeks, just like steam locomotives or vintage cars are today. Computing however will surround us, in Phones in TVs and in many other areas, we may not recognise it though. There will still be workstations, the descendants of today’s desktop PCs. The alternative computer manufacturers may become the only ones left serving those who want the “real thing”.

Before we get to that stage an awful lot is going to happen.
In the short term the current manufacturers and technology leaders will continue to duke it out for a bigger share of the market. Wireless will become the norm but I don’t expect it to have a smooth ride, people are paranoid enough about mobile phones damaging health and expect more powerful devices to cause more controversy. RFID tags will make it into everything but I can see sales of amateur radios going up as the more paranoid try to fry them. Longer term technology trends mean what we know as computers today will change over time into things we may not even consider as computers. Technology will get faster and better, easier to use, smaller and eventually as Moore’s law slows down it’ll last longer as well.

How we build technology will change and an infant industry will become another part of society subject to it’s rules and regulations, it’ll never be the same as the last 30 years, but this will be a gradual process. That’s not to say innovation will die, there’s a lot of technologies not yet our desktops which have yet to play their part. We have seen radical platforms arrive in the 70s and 80s and evolution playing it’s part in the 90s and 00s. I think radicalism will return to the desktop but I don’t know who will have the resources or for that fact, the balls to do it.

The Next Platform

The following is a description of a fictional platform. Like the radical platforms in the 80s it’s based on the combination of mostly existing technologies into something better than anything that has gone before. I don’t know if anyone will build it but I expect any radical new platform which comes along will incorporate at least some of the features from it:

It’ll have a CPU, GPU and some smaller special purpose devices (start with what works first).

It’ll have a highly stable system in the bottom OS layers and at the top it’ll be advanced and user friendly (for the software as well).

The GPU will be more general purpose so it’ll speed things up amazingly when programmed properly (Perhaps a Cell processor?).

There’ll be an FPGA and a collection of libraries so you can use them to boost performance of some intensive operations onto another planet.

It’ll run a hardware virtualising layer so you can run multiple *platforms* simultaneously.

It’ll run anything you want as it’ll include multiple CPU emulators and will to all intents and purposes be ISA agnostic.

The CPU will have message passing, process caching and switching in hardware so there’ll be no performance loss from the micro-kernel OS, in fact these features may mean macro-kernels will be slower.

The GUI will be 3D but work in 2D as well, It’ll be ready for volumetric displays when they become affordable. When they do expect to see a lot of people looking very silly as they wave their hands in the air, the mouse will then become obsolete.

It’ll be really easy to program.

It will include a phone and you will be able to fit it into your pocket (though maybe not in the first version).

And Finally…

That is my view of the Future of Computing and the possibilities it will bring. I don’t expect I’ve been right in everything but no one trying to predict the future ever is. I guess we’ll find out some day.
I hope you’ve enjoyed reading my thoughts.

Thanks to the people who wrote comments and sent me e-mails, there were some very good comments and interesting links to follow.

So, I’ve enjoyed my stint as an anti-historian. What do you expect will happen? Maybe your predictions of the future are completely different, why not write them down, I look forward to reading them.

—————————-

References

[1] A page on AI in Science fiction (Warning: may require sunglasses).

http://www.aaai.org/AITopics/html/scifi.html

[2] Some of Roger Penrose’s thoughts on AI.

http://plus.maths.org/issue18/features/penrose/

[3] AI the Movie.

http://www.indelibleinc.com/kubrick/films/ai/

[4] Yours truly, 2095 from the ELO album “Time”. By Jeff Lynne

http://janbraum.unas.cz/elo/ELO/diskografie/Time.htm

[5] H G Wells’ The War of the Worlds

http://www.bartleby.com/1002/

Jeff Wayne’s musical version of the story is one of my favourite CDs:

http://www.waroftheworldsonline.com/musical.htm

Copyright (c) Nicholas Blachford March 2004

Disclaimer:

This series is about the future and as such is nothing more than informed speculation on my part. I suggest future possibilities and actions which companies may take but this does not mean that they will take them or are even considering them.

50 Comments

  1. 2004-03-11 8:05 pm
  2. 2004-03-11 8:20 pm
  3. 2004-03-11 8:54 pm
  4. 2004-03-11 9:40 pm
  5. 2004-03-11 9:43 pm
  6. 2004-03-11 10:13 pm
  7. 2004-03-11 10:31 pm
  8. 2004-03-11 10:32 pm
  9. 2004-03-11 10:58 pm
  10. 2004-03-11 11:21 pm
  11. 2004-03-11 11:50 pm
  12. 2004-03-11 11:57 pm
  13. 2004-03-12 12:22 am
  14. 2004-03-12 12:22 am
  15. 2004-03-12 12:24 am
  16. 2004-03-12 1:02 am
  17. 2004-03-12 1:07 am
  18. 2004-03-12 1:07 am
  19. 2004-03-12 1:17 am
  20. 2004-03-12 1:21 am
  21. 2004-03-12 1:34 am
  22. 2004-03-12 1:53 am
  23. 2004-03-12 1:53 am
  24. 2004-03-12 1:55 am
  25. 2004-03-12 2:03 am
  26. 2004-03-12 2:17 am
  27. 2004-03-12 2:25 am
  28. 2004-03-12 2:33 am
  29. 2004-03-12 2:42 am
  30. 2004-03-12 2:57 am
  31. 2004-03-12 3:09 am
  32. 2004-03-12 3:28 am
  33. 2004-03-12 3:42 am
  34. 2004-03-12 4:58 am
  35. 2004-03-12 5:11 am
  36. 2004-03-12 5:36 am
  37. 2004-03-12 5:43 am
  38. 2004-03-12 10:38 am
  39. 2004-03-12 10:57 am
  40. 2004-03-12 11:42 am
  41. 2004-03-12 4:17 pm
  42. 2004-03-12 4:25 pm
  43. 2004-03-12 5:08 pm
  44. 2004-03-12 6:41 pm
  45. 2004-03-12 6:45 pm
  46. 2004-03-13 4:58 am
  47. 2004-03-13 8:08 am
  48. 2004-03-15 6:55 am
  49. 2004-03-15 9:29 am