Fujitsu asked BT's former Futurologist, Ian Pearson, to write a report (opens in new tab) on how technology is influencing our lives. Having worked for 16 years at BT peering through his virtual crystal ball, Mr Pearson is ideally placed to track and predict new developments throughout information technology, considering both technological and social implication. He kindly accepted to answer a few of our questions regarding the future of mankind, security, robots and of course, the Matrix and Blade Runner.
1. Mr Pearson, your 'Life and How We'll Live It' Futurizon report talks about smart dust speckles. How far are we from this scenario?
Smart dust already exists; the smallest chips only a fraction of a millimetre across. But with feature sizes still a long way from the limits, they can get a lot smaller. Using the third dimension with many layers, it will be possible to get truly dust sized particles with significant power, storage and sensory capability.
2. The recent case of the MI6 employee who managed to steal tens of thousands of files using a tiny USB drive shows the advances of miniaturisation, but where's the limit?
Engineers keep breaking what used to be considered the limits, but I guess it is hard to imagine getting much further than a few bits per atom, using electrons as a 'Pauli switch' and maybe their spin states also as storage.
Sounds science fiction today, but one day it will be possible. Today, tens of thousands of atoms are used for each bit of storage or each transistor. Another way of looking at it is that all the information stored in 10,000 human brains would fit comfortably in a pin head at such densities.
3. The document you prepare mentions the pervasiveness of the Cloud, something that through aggressive redundancy processes could become autonomous. What impact will this have on people's lives?
Autonomous is an ambiguous term. On one interpretation it can simply mean highly automated. We will certainly be able to make self-organising clouds that provide massive processing power, communications, storage and sensing, all with minimal intervention. They could even self heal and re-route and be hard to crash.
That would make life much easier, but also could eliminate the need for many jobs. But the opportunities created by such a cloud would generate far more, so the balance would be an increase in jobs. We are all used to the idea of continuous learning and the need to adapt. This just takes it to the next level.
The other interpretation suggests consciousness. That is also feasible, but not with the kinds of chips and IT we have today. I strongly believe that machine consciousness will require a high degree of analogue computing. Adaptive analogue neurons are the way to go. Carver Mead said that around 1990.
He was right then and nothing that has happened since has changed my mind that that is the only approach that will work. I don't think we need to worry about self awareness and consciousness arising from digital IT. It is too fragile, too complex and at best can only ever offer a simulation of consciousness, never emulation.
4. Back in 2002, you told the BBC about the possibility of getting a robot to pass its GCSE exams. When do you think that will happen now?
Actually I am surprised no-one has bothered to do this yet. We already have all the stuff needed. The character recognition, the easy access to maths programmes such as Mathematica, or knowledge databases for other subjects, the ability to print or even make a robot arm use a pencil.
I think any university IT department could knock up a GCSE-passing robot in a few days now for a publicity stunt. They just haven't bothered. Maybe one will next year or the year after. Certainly some expert systems out there in law, medicine and science perform at a far higher level than is required for GCSE.
5. Your research points to the fact that remote working will become more popular because of the ubiquity of cloud computing. But then, doesn't that mean that everything that can be outsourced or carried out remotely, can also be automated and carried out by machines?
No, not at all. Machines and people have different skill sets. Much of the work that people do will be automated for sure, but there will remain a lot that can't, and it will still be done by people. I call it the care economy, when the information economy has been reduced to low cost by machine automation, and human, interpersonal, emotional skills are the main focus of work. Making machines smarter just forces people to become more human. That's a good thing. Even the remote working will only work for part of our work then, much of our work will be personal services and will need to be delivered face to face.
6. When do you think the Internet (or machines) will become sentient? Will the Matrix ever become a reality?
I have argued since 1990 that the first conscious machines will become reality between 2015 and 2020. Most scientists think that is too early but I still think it is correct. You need to factor in positive feedback. Better computers are helping us design better tools for brains studies, which are yielding more insights into the workings of cognition.
Taking some insights as input, using fast processing to mess around with them, adding evolutionary software and hardware, and all the while neuroscientists are talking to IT people - we will see progress in the adaptive analogue neurons we need for consciousness, and the whole thing will accelerate towards the goal. Almost all the progress will happen in the last 2 years, just as it did with the human genome project, so people will still be laughing at how ridiculous the whole idea is until the year before it happens.
The Matrix was a poor story. If the computers are networked and smart enough to capture our minds like that, we would long since have gone through a level where we could network all our minds together and produce a global consciousness - like the Borg but without the stupidity. All people and all machines could be networked. That is much more fun idea-wise. Mental warfare among factions of a global consciousness would have made a better film, instead of gratuitous martial arts and slow motion graphics.
7. Many new advances try to better humanity (lengthening lifespan, curing illnesses, boosting physical power, increasing brain capacity). But none of them make people intrinsically better. Is there anything in your crystal ball that will make human beings better?
We can certainly make humans technically better, with better senses, higher intelligence, greater sports prowess, longer life and so on. The trouble with 'better' generally though is we don't agree what it means, and that is the weakness. It implies tweaking human nature, making us wiser instead of just smarter.
I wrote a paper on it called 'Wisdom' on the Futurizon site. After a long argument I concluded that we can't really do it. All we can hope for is temporary improvements in wisdom. So, no, I can't see anything coming over the horizon that can make us better.
8. What aspects of life do you think should get more technological investments and why?
If I was offered huge funding to research what I wanted to improve our lot, I would put almost all of it into AI. If we can get much smarter tools than our feeble brains, we could develop anything else far more easily. It's as simple as that. Give me a computer with an effective IQ of a million, we can solve many of mankind's problems before breakfast.
9. Why haven't we seen more Cyborgs roaming around à la Blade Runner?
Well, we can't build anything like them yet. We have polymer gels, primitive robotics, some basic genetic tools, and some low quality AI. But that's not a bad start. Give it another two decades and we will have the gels to make the muscles at low cost, the robotics will allow them to walk around and do basic tasks, the sex-toy manufacturers will be able to make them look nice, and the AI will be capable of pretty good conversation, probably conscious and probably as smart or smarter than us. The idea that we would use GM to make them is interesting, rather than simple robotic build. The jury is out on that, I think it could go either way. Each path has a lot of merit.