Optical computing

A few nights ago I was thinking about the optical fibre memories that we were designing in the late 1980s in BT. The idea was simple. You transmit data into an optical fibre, and if the data rate is high you can squeeze lots of data into a manageable length. Back then the speed of light in fibre was about 5 microseconds per km of fibre, so 1000km of fibre, at a data rate of 2Gb/s would hold 10Mbits of data, per wavelength, so if you can multiplex 2 million wavelengths, you’d store 20Tbits of data. You could maintain the data by using a repeater to repeat the data as it reaches one end into the other, or modify it at that point simply by changing what you re-transmit. That was all theory then, because the latest ‘hero’ experiments were only just starting to demonstrate the feasibility of such long lengths, such high density WDM and such data rates.

Nowadays, that’s ancient history of course, but we also have many new types of fibre, such as hollow fibre with various shaped pores and various dopings to allow a range of effects. And that’s where using it for computing comes in.

If optical fibre is designed for this purpose, with optimal variable refractive index designed to facilitate and maximise non-linear effects, then the photons in one data stream on one wavelength could have enough effects of photons in another stream to be used for computational interaction. Computers don’t have to be digital of course, so the effects don’t have to be huge. Analog computing has many uses, and analog interactions could certainly work, while digital ones might work, and hybrid digital/analog computing may also be feasible. Then it gets fun!

Some of the data streams could be programs. Around that time, I was designing protocols with smart packets that contained executable code, as well as other packets that could hold analog or digital data or any mix. We later called the smart packets ANTs – autonomous network telephers, a contrived term if ever there was one, but we wanted to call them ants badly. They would scurry around the network doing a wide range of jobs, using a range of biomimetic and basic physics techniques to work like ant colonies and achieve complex tasks using simple means.

If some of these smart packets or ANTs are running along a fibre, changing the properties as they go to interact with other data transmitting alongside, then ANTs can interact with one another and with any stored data. ANTs could also move forwards or backwards along the fibre by using ‘sidings’ or physical shortcuts, since they can route themselves or each other. Data produced or changed by the interactions could be digital or analog and still work fine, carried on the smart packet structure.

(If you’re interested my protocol was called UNICORN, Universal Carrier for an Optical Residential Network, and used the same architectural principles as my previous Addressed Time Slice invention, compressing analog data by a few percent to fit into a packet, with a digital address and header, or allowing any digital data rate or structure in a payload while keeping the same header specs for easy routing. That system was invented (in 1988) for the late 1990s when basic domestic broadband rate should have been 625Mbit/s or more, but we expected to be at 2Gbit/s or even 20Gbit/s soon after that in the early 2000s, and the benefit as that we wouldn’t have to change the network switching because the header overheads would still only be a few percent of total time. None of that happened because of government interference in the telecoms industry regulation that strongly disincentivised its development, and even today, 625Mbit/s ‘basic rate’ access is still a dream, let alone 20Gbit/s.)

Such a system would be feasible. Shortcuts and sidings are easy to arrange. The protocols would work fine. Non-linear effects are already well known and diverse. If it were only used for digital computing, it would have little advantage over conventional computers. With data stored on long fibre lengths, external interactions would be limited, with long latency. However, it does present a range of potentials for use with external sensors directly interacting with data streams and ANTs to accomplish some tasks associated with modern AI. It ought to be possible to use these techniques to build the adaptive analog neural networks that we’ve known are the best hope of achieving strong AI since Hans Moravek’s insight, coincidentally also around that time. The non-linear effects even enable ideal mechanisms for implementing emotions, biasing the computation in particular directions via intensity of certain wavelengths of light in much the same way as chemical hormones and neurotransmitters interact with our own neurons. Implementing up to 2 million different emotions at once is feasible.

So there’s a whole mineful of architectures, tools and techniques waiting to be explored and mined by smart young minds in the IT industry, using custom non-linear optical fibres for optical AI.

AI could use killer drone swarms to attack people while taking out networks

In 1987 I discovered a whole class of security attacks that could knock out networks, which I called correlated traffic attacks, creating particular patterns of data packet arrivals from particular sources at particular times or intervals. We simulated two examples to successfully verify the problem. One example was protocol resonance. I demonstrated that it was possible to push a system into a gross overload state with a single call, by spacing the packets precise intervals apart. Their arrival caused a strong resonance in the bandwidth allocation algorithms and the result was that network capacity was instantaneously reduced by around 70%. Another example was information waves, whereby a single piece of information appearing at a particular point could, by its interaction with particular apps on mobile devices (the assumption was financially relevant data that would trigger AI on the devices to start requesting voluminous data, triggering a highly correlated wave of responses, using up bandwidth and throwing the network into overload, very likely crashing due to initiation of rarely used software. When calls couldn’t get through, the devices would wait until the network recovered, then they would all simultaneously detect recovery and simultaneously try again, killing the net again, and again, until people were asked to turn  their devices off and on again, thereby bringing randomness back into the system. Both of these examples could knock out certain kinds of networks, but they are just two of an infinite set of possibilities in the correlated traffic attack class.

Adversarial AI pits one AI against another, trying things at random or making small modifications until a particular situation is achieved, such as the second AI accepting an image is acceptable. It is possible, though I don’t believe it has been achieved yet, to use the technique to simulate a wide range of correlated traffic situations, seeing which ones achieve network resonance or overloads, which trigger particular desired responses from network management or control systems, via interactions with the network and its protocols, commonly resident apps on mobile devices or computer operating systems.

Activists and researchers are already well aware that adversarial AI can be used to find vulnerabilities in face recognition systems and thereby prevent recognition, or to deceive autonomous car AI into seeing fantasy objects or not seeing real ones. As Noel Sharkey, the robotics expert, has been tweeting today, it will be possible to use adversarial AI to corrupt recognition systems used by killer drones, potentially to cause them to attack their controllers or innocents instead of their intended targets. I have to agree with him. But linking that corruption to the whole extended field of correlated traffic attacks extends the range of mechanisms that can be used greatly. It will be possible to exploit highly obscured interactions between network physical architecture, protocols and operating systems, network management, app interactions, and the entire sensor/IoT ecosystem, as well as software and AI systems using it. It is impossible to check all possible interactions, so no absolute defence is possible, but adversarial AI with enough compute power could randomly explore across these multiple dimensions, stumble across regions of vulnerability and drill down until grand vulnerabilities are found.

This could further be linked to apps used as highly invisible Trojans, offering high attractiveness to users with no apparent side effects, quietly gathering data to help identify potential targets, and simply waiting for a particular situation or command before signalling to the attacking system.

A future activist or terrorist group or rogue state could use such tools to make a multidimensional attack. It could initiate an attack, using its own apps to identify and locate targets, control large swarms of killer drones or robots to attack them, simultaneously executing a cyberattack that knocks out selected parts of the network, crashing or killing computers and infrastructure. The vast bulk of this could be developed, tested and refined offline, using simulation and adversarial AI approaches to discover vulnerabilities and optimise exploits.

There is already debate about killer drones, mainly whether we should permit them and in what circumstances, but activists and rogue states won’t care about rules. Millions of engineers are technically able to build such things and some are not on your side. It is reasonable to expect that freely available AI tools will be used in such ways, using their intelligence to design, refine, initiate and control attacks using killer drones, robots and self-driving cars to harm us, while corrupting systems and infrastructure that protect us.

Worrying, especially since the capability is arriving just as everyone is starting to consider civil war.

 

 

Some trees just don’t get barked up

Now and then, someone asks me for an old document and as I search for it, I stumble across others I’d forgotten about. I’ve been rather frustrated that AI progress hasn’t kept up with its development rate in the 90s, so this was fun to rediscover, highlighting some future computing directions that offered serious but uncertain potential exactly 20 years ago, well 20 years ago 3 weeks ago. Here is the text, and the Schrodinger’s Computer was only ever intended to be silly (since renamed the Yonck Processor):

Herrings, a large subset of which are probably red

Computers in the future will use a wide range of techniques, not just conventional microprocessors. Problems should be decomposed and the various components streamed to the appropriate processing engines. One of the important requirements is therefore some means of identifying automatically which parts of a problem could best be tackled by which techniques, though sometimes it might be best to use several in parallel with some interaction between them.

 Analogs

We have a wider variety of components available to be used in analog computing today than we had when it effectively died out in the 80s. With much higher quality analog and mixed components, and additionally micro-sensors, MEMs, simple neural network components, and some imminent molecular capability, how can we rekindle the successes of the analog domain. Nature handles the infinite body problem with ease! Things just happen according to the laws of physics. How can we harness them too? Can we build environments with synthetic physics to achieve more effects? The whole field of non-algorithmic computation seems ripe for exploitation.

 Neural networks

• Could we make neural microprocessor suspensions, using spherical chips suspended in gel in a reflective capsule and optical broadcasting. Couple this with growing wires across the electric field. This would give us both electrical and optical interconnection that could be ideal for neural networks with high connectivity. Could link this to gene chip technology to have chemical detection and synthesis on the chips too, so that we could have close high speed replicas of organic neural networks.
• If we can have quantum entanglement between particles, might this affect the way in which neurons in the brain work? Do we have neural entanglement and has this anything to do with how our brain works. Could we create neural entanglement or even virtual entanglement and would it have any use?
• Could we make molecular neurons (or similar) using ordinary chemistry? And then form them into networks. Might need nanomachines and bottom-up assembly.
• Could we use neurons as the first stage filters to narrow down the field to make problems tractable for other techniques
• Optical neurons
• Magnetic neurons

Electromechanical, MEMS etc

• Micromirror arrays as part of optical computers, perhaps either as data entry, or as part of the algorithm
• Carbon fullerene balls and tubes as MEM components
• External fullerene ‘décor’ as a form of information, cf antibodies in immune system
• Sensor suspensions and gels as analog computers for direct simulation

Interconnects

• Carbon fullerene tubes as on chip wires
• Could they act as electron pipes for ultra-high speed interconnect
• Optical or radio beacons on chip

Software

• Transforms – create a transform of every logic component, spreading the functionality across a wide domain, and construct programs using them instead. Small perturbation is no longer fatal but just reduces accuracy
• Filters – nature works often using simple physical effects where humans design complex software. We need to look at hard problems to see how we might make simple filters to narrow the field before computing final details and stages conventionally.
• Interference – is there some form of representation that allows us to compute operations by means of allowing the input data to interact directly, i.e. interference, instead of using tedious linear computation. Obviously only suited to a subset of problems.

And finally, the frivolous

• Schrodinger’s computer – design of computer and software, if any, not determined until box is opened. The one constant is that it destroys itself if it doesn’t finding the solution. All possible computers and all possible programs exist and if there is a solution, the computer will pop out alive and well with the answer. Set it the problem of answering all possible questions too, working out which ones have the most valuable answers and using up all the available storage to write the best answers.

The future of reproductive choice

I’m not taking sides on the abortion debate, just drawing maps of the potential future, so don’t shoot the messenger.

An average baby girl is born with a million eggs, still has 300,000 when she reaches puberty, and subsequently releases 300 – 400 of these over her reproductive lifetime. Typically one or two will become kids but today a woman has no way of deciding which ones, and she certainly has no control over which sperm is used beyond choosing her partner.

Surely it can’t be very far in the future (as a wild guess, say 2050) before we fully understand the links between how someone is and their genetics (and all the other biological factors involved in determining outcome too). That knowledge could then notionally be used to create some sort of nanotech (aka magic) gate that would allow her to choose which of her eggs get to be ovulated and potentially fertilized, wasting ones she isn’t interested in and going for it when she’s released a good one. Maybe by 2060, women would also be able to filter sperm the same way, helping some while blocking others. Choice needn’t be limited to whether to have a baby or not, but which baby.

Choosing a particularly promising egg and then which sperm would combine best with it, an embryo might be created only if it is likely to result in the right person (perhaps an excellent athlete, or an artist, or a scientist, or just good looking), or deselected if it would become the wrong person (e.g. a terrorist, criminal, saxophonist, Republican).

However, by the time we have the technology to do that, and even before we fully know what gene combos result in what features, we would almost certainly be able to simply assemble any chosen DNA and insert it into an egg from which the DNA has been removed. That would seem a more reliable mechanism to get the ‘perfect’ baby than choosing from a long list of imperfect ones. Active assembly should beat deselection from a random list.

By then, we might even be using new DNA bases that don’t exist in nature, invented by people or AI to add or control features or abilities nature doesn’t reliably provide for.

If we can do that, and if we know how to simulate how someone might turn out, then we could go further and create lots of electronic babies that live their entire lives in an electronic Matrix style existence. Let’s expand on that briefly.

Even today, couples can store eggs and sperm for later use, but with this future genetic assembly, it will become feasible to create offspring from nothing more than a DNA listing. DNA from both members of a couple, of any sex, could get a record of their DNA, randomize combinations with their partner’s DNA and thus get a massive library of potential offspring. They may even be able to buy listings of celebrity DNA from the net. This creates the potential for greatly delayed birth and tradable ‘ebaybies’ – DNA listings are not alive so current laws don’t forbid trading in them. These listings could however be used to create electronic ‘virtual’offspring, simulated in a computer memory instead of being born organically. Various degrees of existence are possible with varied awareness. Couples may have many electronic babies as well as a few real ones. They may even wait to see how a simulation works out before deciding which kids to make for real. If an electronic baby turns out particularly well, it might be promoted to actual life via DNA assembly and real pregnancy. The following consequences are obvious:

Trade-in and collection of DNA listings, virtual embryos, virtual kids etc, that could actually be fabricated at some stage

Re-birth, potential to clone and download one’s mind or use a direct brain link to live in a younger self

Demands by infertile and gay couples to have babies via genetic assembly

Ability of kids to own entire populations of virtual people, who are quite real in some ways.

It is clear that this whole technology field is rich in ethical issues! But we don’t need to go deep into future tech to find more of those. Just following current political trends to their logical conclusions introduces a lot more. I’ve written often on the random walk of values, and we cannot be confident that many values we hold today will still reign in decades time. Where might this random walk lead? Let’s explore some more.

Even in ‘conventional’ pregnancies, although the right to choose has been firmly established in most of the developed world, a woman usually has very little information about the fetus and has to make her decision almost entirely based on her own circumstances and values. The proportion of abortions related to known fetal characteristics such as genetic conditions or abnormalities is small. Most decisions can’t yet take any account of what sort of person that fetus might become. We should expect future technology to provide far more information on fetus characteristics and likely future development. Perhaps if a woman is better informed on likely outcomes, might that sometimes affect her decision, in either direction?

In some circumstances, potential outcome may be less certain and an informed decision might require more time or more tests. To allow for that without reducing the right to choose, is possible future law could allow for conditional terminations, registered before a legal time limit but performed later (before another time limit) when more is known. This period could be used for more medical tests, or to advertise the baby to potential adopters that want a child just like that one, or simply to allow more time for the mother to see how her own circumstances change. Between 2005 and 2015, USA abortion rate dropped from 1 in 6 pregnancies to 1 in 7, while in the UK, 22% of pregnancies are terminated. What would these figures be if women could determine what future person would result? Would termination rate increase? To 30%, 50%? Abandon this one and see if we can make a better one? How many of us would exist if our parents had known then what they know now?

Whether and how late terminations should be permitted is still fiercely debated. There is already discussion about allowing terminations right up to birth and even after birth in particular circumstances. If so, then why stop there? We all know people who make excellent arguments for retrospective abortion. Maybe future parents should be allowed to decide whether to keep a child right up until it reaches its teens, depending on how the child turns out. Why not 16, or 18, or even 25, when people truly reach adulthood? By then they’d know what kind of person they’re inflicting on the world. Childhood and teen years could simply be a trial period. And why should only the parents have a say? Given an overpopulated world with an infinite number of potential people that could be brought into existence, perhaps the state could also demand a high standard of social performance before assigning a life license. The Chinese state already uses surveillance technology to assign social scores. It is a relatively small logical step further to link that to life licenses that require periodic renewal. Go a bit further if you will, and link that thought to the blog I just wrote on future surveillance: https://timeguide.wordpress.com/2019/05/19/future-surveillance/.

Those of you who have watched Logan’s Run will be familiar with the idea of  compulsory termination at a certain age. Why not instead have a flexible age that depends on social score? It could range from zero to 100. A pregnancy might only be permitted if the genetic blueprint passes a suitability test and then as nurture and environmental factors play their roles as a person ages, their life license could be renewed (or not) every year. A range of crimes might also result in withdrawal of a license, and subsequent termination.

Finally, what about AI? Future technology will allow us to make hybrids, symbionts if you like, with a genetically edited human-ish body, and a mind that is part human, part AI, with the AI acting partly as enhancement and partly as a control system. Maybe the future state could insist that installation into the embryo of a state ‘guardian’, a ‘supervisory AI’, essentially a deeply embedded police officer/judge/jury/executioner will be required to get the life license.

Random walks are dangerous. You can end up where you start, or somewhere very far away in any direction.

The legal battles and arguments around ‘choice’ won’t go away any time soon. They will become broader, more complex, more difficult, and more controversial.

Future Surveillance

This is an update of my last surveillance blog 6 years ago, much of which is common discussion now. I’ll briefly repeat key points to save you reading it.

They used to say

“Don’t think it

If you must think it, don’t say it

If you must say it, don’t write it

If you must write it, don’t sign it”

Sadly this wisdom is already as obsolete as Asimov’s Laws of Robotics. The last three lines have already been automated.

I recently read of new headphones designed to recognize thoughts so they know what you want to listen to. Simple thought recognition in various forms has been around for 20 years now. It is slowly improving but with smart networked earphones we’re already providing an easy platform into which to sneak better monitoring and better though detection. Sold on convenience and ease of use of course.

You already know that Google and various other large companies have very extensive records documenting many areas of your life. It’s reasonable to assume that any or all of this could be demanded by a future government. I trust Google and the rest to a point, but not a very distant one.

Your phone, TV, Alexa, or even your networked coffee machine may listen in to everything you say, sending audio records to cloud servers for analysis, and you only have naivety as defense against those audio records being stored and potentially used for nefarious purposes.

Some next generation games machines will have 3D scanners and UHD cameras that can even see blood flow in your skin. If these are hacked or left switched on – and social networking video is one of the applications they are aiming to capture, so they’ll be on often – someone could watch you all evening, capture the most intimate body details, film your facial expressions and gaze direction while you are looking at a known image on a particular part of the screen. Monitoring pupil dilation, smiles, anguished expressions etc could provide a lot of evidence for your emotional state, with a detailed record of what you were watching and doing at exactly that moment, with whom. By monitoring blood flow and pulse via your Fitbit or smartwatch, and additionally monitoring skin conductivity, your level of excitement, stress or relaxation can easily be inferred. If given to the authorities, this sort of data might be useful to identify pedophiles or murderers, by seeing which men are excited by seeing kids on TV or those who get pleasure from violent games, and it is likely that that will be one of the justifications authorities will use for its use.

Millimetre wave scanning was once controversial when it was introduced in airport body scanners, but we have had no choice but to accept it and its associated abuses –  the only alternative is not to fly. 5G uses millimeter wave too, and it’s reasonable to expect that the same people who can already monitor your movements in your home simply by analyzing your wi-fi signals will be able to do a lot better by analyzing 5G signals.

As mm-wave systems develop, they could become much more widespread so burglars and voyeurs might start using them to check if there is anything worth stealing or videoing. Maybe some search company making visual street maps might ‘accidentally’ capture a detailed 3d map of the inside of your house when they come round as well or instead of everything they could access via your wireless LAN.

Add to this the ability to use drones to get close without being noticed. Drones can be very small, fly themselves and automatically survey an area using broad sections of the electromagnetic spectrum.

NFC bank and credit cards not only present risks of theft, but also the added ability to track what we spend, where, on what, with whom. NFC capability in your phone makes some parts of life easier, but NFC has always been yet another doorway that may be left unlocked by security holes in operating systems or apps and apps themselves carry many assorted risks. Many apps ask for far more permissions than they need to do their professed tasks, and their owners collect vast quantities of information for purposes known only to them and their clients. Obviously data can be collected using a variety of apps, and that data linked together at its destination. They are not all honest providers, and apps are still very inadequately regulated and policed.

We’re seeing increasing experimentation with facial recognition technology around the world, from China to the UK, and only a few authorities so far such as in San Francisco have had the wisdom to ban its use. Heavy handed UK police, who increasingly police according to their own political agenda even at the expense of policing actual UK law, have already fined people who have covered themselves to avoid being abused in face recognition trials. It is reasonable to assume they would gleefully seize any future opportunity to access and cross-link all of the various data pools currently being assembled under the excuse of reducing crime, but with the real intent of policing their own social engineering preferences. Using advanced AI to mine zillions of hours of full-sensory data input on every one of us gathered via all this routine IT exposure and extensive and ubiquitous video surveillance, they could deduce everyone’s attitudes to just about everything – the real truth about our attitudes to every friend and family member or TV celebrity or politician or product, our detailed sexual orientation, any fetishes or perversions, our racial attitudes, political allegiances, attitudes to almost every topic ever aired on TV or everyday conversation, how hard we are working, how much stress we are experiencing, many aspects of our medical state.

It doesn’t even stop with public cameras. Innumerable cameras and microphones on phones, visors, and high street private surveillance will automatically record all this same stuff for everyone, sometimes with benign declared intentions such as making self-driving vehicles safer, sometimes using social media tribes to capture any kind of evidence against ‘the other’. In depth evidence will become available to back up prosecutions of crimes that today would not even be noticed. Computers that can retrospectively date mine evidence collected over decades and link it all together will be able to identify billions of real or invented crimes.

Active skin will one day link your nervous system to your IT, allowing you to record and replay sensations. You will never be able to be sure that you are the only one that can access that data either. I could easily hide algorithms in a chip or program that only I know about, that no amount of testing or inspection could ever reveal. If I can, any decent software engineer can too. That’s the main reason I have never trusted my IT – I am quite nice but I would probably be tempted to put in some secret stuff on any IT I designed. Just because I could and could almost certainly get away with it. If someone was making electronics to link to your nervous system, they’d probably be at least tempted to put a back door in too, or be told to by the authorities.

The current panic about face recognition is justified. Other AI can lipread better than people and recognize gestures and facial expressions better than people. It adds the knowledge of everywhere you go, everyone you meet, everything you do, everything you say and even every emotional reaction to all of that to all the other knowledge gathered online or by your mobile, fitness band, electronic jewelry or other accessories.

Fools utter the old line: “if you are innocent, you have nothing to fear”. Do you know anyone who is innocent? Of everything? Who has never ever done or even thought anything even a little bit wrong? Who has never wanted to do anything nasty to anyone for any reason ever? And that’s before you even start to factor in corruption of the police or mistakes or being framed or dumb juries or secret courts. The real problem here is not the abuses we already see. It is what is being and will be collected and stored, forever, that will be available to all future governments of all persuasions and police authorities who consider themselves better than the law. I’ve said often that our governments are often incompetent but rarely malicious. Most of our leaders are nice guys, only a few are corrupt, but most are technologically inept . With an increasingly divided society, there’s a strong chance that the ‘wrong’ government or even a dictatorship could get in. Which of us can be sure we won’t be up against the wall one day?

We’ve already lost the battle to defend privacy. The only bits left are where the technology hasn’t caught up yet. In the future, not even the deepest, most hidden parts of your mind will be private. Pretty much everything about you will be available to an AI-upskilled state and its police.

The future for women, pdf version

It is several years since my last post on the future as it will affect women so here is my new version as a pdf presentation:

Women and the Future

Augmented reality will objectify women

Microsoft Hololens 2 Visor

The excitement around augmented reality continues to build, and I am normally  enthusiastic about its potential, looking forward to enjoying virtual architecture, playing immersive computer games, or enjoying visual and performance artworks transposed into my view of the high street while I shop.

But it won’t all be wonderful. While a few PR and marketing types may worry a little about people overlaying or modifying their hard-won logos and ads, a bigger issue will be some people choosing to overlay people in the high street with ones that are a different age or gender or race, or simply prettier. Identity politics will be fought on yet another frontier.

In spite of waves of marketing hype and misrepresentation, AR is really only here in primitive form outside the lab. Visors fall very far short of what we’d hoped for by now even a decade ago, even the Hololens 2 shown above. But soon AR visors and eventually active contact lenses will enable fully 3D hi-res overlays on the real world. Then, in principle at least, you can make things look how you want, with a few basic limits. You could certainly transform a dull shop, cheap hotel room or an office into an elaborate palace or make it look like a spaceship. But even if you change what things look like, you still have to represent nearby physical structures and obstacles in your fantasy overlay world, or you may bump into them, and that includes all the walls and furniture, lamp posts, bollards, vehicles, and of course other people. Augmented reality allows you to change their appearance thoroughly but they still need to be there somehow.

When it comes to people, there will be some battles. You may spend ages creating a wide variety of avatars, or may invest a great deal of time and money making or buying them. You may have a digital aura, hoping to present different avatars to different passers-by according to their profiles. You may want to look younger or thinner or as a character you enjoy playing in a computer game. You may present a selection of options to the AIs controlling the passer person’s view and the avatar they see overlaid could be any one of the images you have on offer. Perhaps some privileged people get to pick from a selection you offer, while others you wish to privilege less are restricted to just one that you have set for their profile. Maybe you’d have a particularly ugly or offensive one to present to those with opposing political views.

Except that you can’t assume you will be in control. In fact, you probably won’t.

Other people may choose not to see your avatar, but instead to superimpose one of their own choosing. The question of who decides what the viewer sees is perhaps the first and most important battle in AR. Various parties would like to control it – visor manufacturers, O/S providers, UX designers, service providers, app creators, AI providers, governments, local councils, police and other emergency services, advertisers and of course individual users. Given market dynamics, most of these ultimately come down to user choice most of the time, albeit sometimes after paying for the privilege. So it probably won’t be you who gets to choose how others see you, via assorted paid intermediary services, apps and AI, it will be the other person deciding how they want to see you, regardless of your preferences.

So you can spend all the time you want designing your avatar and tweaking your virtual make-up to perfection, but if someone wants to see their favorite celebrity walking past instead of you, they will. You and your body become no more than an object on which to display any avatar or image someone else chooses. You are quite literally reduced to an object in the AR world. Augmented reality will literally objectify women, reducing them to no more than a moving display space onto which their own selected images are overlaid. A few options become obvious.

Firstly they may just take your actual physical appearance (via a video camera built into their visor for example) and digitally change it,  so it is still definitely you, but now dressed more nicely, or dressed in sexy lingerie, or how you might look naked, using the latest AI to body-fit fantasy images from a porn database. This could easily be done automatically in real time using some app or other. You’ve probably already seen recent AI video fakery demos that can present any celebrity saying anything at all, almost indistinguishable from reality. That will soon be pretty routine tech for AR apps. They could even use your actual face as input to image-matching search engines to find the most plausible naked lookalikes. So anyone could digitally dress or undress you, not just with their eyes, but with a hi-res visor using sophisticated AI-enabled image processing software. They could put you in any kind of outfit, change your skin color or make-up or age or figure, and make you look as pretty and glamorous or as slutty as they want. And you won’t have any idea what they are seeing. You simply won’t know whether they are respectfully celebrating your inherent beauty, or flattering you by making you look even prettier, which you might not mind at all, or might object to strongly in the absence of explicit consent, or worse still, stripping or degrading you to whatever depths they wish, with no consent or notification, which you probably will mind a lot.

Or they can treat you as just an object on which to superimpose some other avatar, which could be anything or anyone – a zombie, favorite actress or supermodel. They won’t need your consent and again you won’t have any idea what they are seeing. The avatar may make the same gestures and movements and even talk plausibly, saying whatever their AI thinks they might like, but it won’t be you. In some ways this might not be so bad. You’d still be reduced to an object but at least it wouldn’t be you that they’re looking at naked. To most strangers on a high street most of the time, you’re just a moving obstacle to avoid bumping into, so being digitally transformed into a walking display board may worry you. Most people will cope with that bit. It is when you stop being just a passing stranger and start to interact in some way that it really starts to matter. You probably won’t like it if someone is chatting to you but they are actually looking at someone else entirely, especially if the viewer is one of your friends or your partner. And if your partner is kissing or cuddling you but seeing someone else, that would be a strong breach of trust, but how would you know? This sort of thing could and probably will damage a lot of relationships.

Most of the software to do most of this is already in development and much is already demonstrable. The rest will develop quickly once AR visors become commonplace.

In the office, in the home, when you’re shopping or at a party, you soon won’t have any idea what or who someone else is seeing when they look at you. Imagine how that would clash with rules that are supposed to be protection from sexual harassment  in the office. Whole new levels of harassment will be enabled, much invisible. How can we police behaviors we can’t even detect? Will hardware manufacturers be forced to build in transparency and continuous experience recording

The main casualty will be trust.  It will make us question how much we trust each of our friends and colleagues and acquaintances. It will build walls. People will often become suspicious of others, not just strangers but friends and colleagues. Some people will become fearful. You may dress as primly or modestly as you like, but if the viewer chooses to see you wearing a sexy outfit, perhaps their behavior and attitude towards you will be governed by that rather than reality. Increased digital objectification might lead to increase physical sexual assault or rape. We may see more people more often objectifying women in more circumstances.

The tech applies equally to men of course. You could make a man look like a silverback gorilla or a zombie or fake-naked. Some men will care more than others, but the vast majority of real victims will undoubtedly be women. Many men objectify women already. In the future AR world , they’ll be able to do so far more effectively, more easily.

 

Who controls AI, controls the world

This week, the fastest supercomputer broke a world record for AI, using machine learning in climate research:

https://www.wired.com/story/worlds-fastest-supercomputer-breaks-ai-record/

I guess most readers thought this is a great thing, after all we need to solve climate change. That wasn’t my thought. The first thing my boss told me when I used a computer for the first time was: “shit in, shit out”. I don’t remember his name but I remember that concise lesson every time I read about climate models. If either the model or the data is garbage, or both, the output will also be garbage.

So my first thought reading about this new record was: will they let the AI work everything out for itself using all the raw, unadjusted data available about the environment, including all the astrophysics data about every kind of solar activity, human agricultural, industrial activities, air travel, all the unadjusted measurements of or proxies for surface, sea and air temperatures, ever collected, any empirical evidence for any corrections that might be needed on such data in any direction, and then let it make its own deductions, form its own models of how it might all connected and then watch eagerly as it makes predictions?

Or will they just input their own models, CO2 blinkering, prejudices and group-think, adjusted datasets, data omissions and general distortions of historical records into biased models already indoctrinated with climate change dogma, so that it will reconfirm the doom and gloom forecasts we’re so used to hearing, maximizing their chances of continued grants? If they do that, the AI might as well be a cardboard box with a pre-written article stuck on it. Shit in, shit out.

It’s obvious that the speed and capability of the supercomputer is of secondary important to who controls the AI, and its access to data, and its freedom to draw its own conclusions.

(Read my blog on Fake AI: https://timeguide.wordpress.com/2017/11/16/fake-ai/)

You may recall a week or two ago that IBM released a new face database to try to address bias in AI face recognition systems. Many other kinds of data could have biases for all sorts of reasons. At face value reducing bias is a good thing, but what exactly do we mean by that? Who decides what is biased and what is real? There are very many potential AI uses that are potentially sensitive, such as identifying criminals or distinguishing traits that correlate with gender, sexuality, race, religion, or indeed any discernible difference. Are all deductions by the AI permissible, or are huge swathes of possible deductions not permitted because they might be politically unacceptable? Who controls the AI? Why? With what aims?

Many people have some degree of influence on  AI. Those who provide funding, equipment, theoreticians, those who design hardware, those who design the learning and training mechanisms, those who supply the data, those who censor or adjust data before letting the AI see it, those who design the interfaces, those who interpret and translate the results, those who decide which results are permissible and how to spin them, and publish them.

People are often impressed when a big powerful computer outputs results of massive amounts of processing. Outputs may often be used to control public opinion and government policy, to change laws, to alter balance of power in society, to create and destroy empires. AI will eventually make or influence most decisions of any consequence.

As AI techniques become more powerful, running on faster and better computers, we must always remember that golden rule: shit in, shit out. And we must always be suspicious of those who might have reason to influence an outcome.

Because who controls AI, controls the world.

 

 

Future AI: Turing multiplexing, air gels, hyper-neural nets

Just in time to make 2018 a bit less unproductive, I managed to wake in the middle of the night with another few inventions. I’m finishing the year on only a third as many as 2016 and 2017, but better than some years. And I quite like these new ones.

Gel computing is a very old idea of mine, and I’m surprised no company has started doing it yet. Air gel is different. My original used a suspension of processing particles in gel, and the idea was that the gel would hold the particles in fixed locations with good free line of sight to neighbor devices for inter-device optical comms, while acting also as a coolant.

Air gel uses the same idea of suspending particles, but does so by using ultrasound, standing waves holding the particles aloft. They would form a semi-gel I suppose, much softer. The intention is that they will be more easily movable than in a gel, and maybe rotate. I imagine using rotating magnetic fields to rotate them, and use that mechanism to implement different configurations of inter-device nets. That would be the first pillar of running multiple neural nets in the same space at the same time, using spin-based TDM (time division multiplexing), or synchronized space multiplexing if you prefer. If a device uses on board processing that is fast compared to the signal transmission time to other devices (the speed of light may be fast but can still be severely limiting for processing and comms), then having the ability to deal with processing associated with several other networks while awaiting a response allows a processing network to be multiplied up several times. A neural net could become a hyper-neural net.

Given that this is intended for mid-century AI, I’m also making the assumption that true TDM can also be used on each net, my second pillar. Signals would carry a stream of slots holding bits for each processing instance. Since this allows a Turing machine to implement many different processes in parallel, I decided to call it Turing multiplexing. Again, it helps alleviate the potential gulf between processing and communication times. Combining Turing and spin multiplexing would allow a single neural net to be multiplied up potentially thousands or millions of times – hyper-neurons seems as good a term as any.

The third pillar of this system is that the processing particles (each could contain a large number of neurons or other IT objects) could be energized and clocked using very high speed alternating EM fields – radio, microwaves, light, even x-rays. I don’t have any suggestions for processing mechanisms that might operate at such frequencies, though Pauli switches might work at lower speeds, using Pauli exclusion principle to link electron spin states to make switches. I believe early versions of spin cubits use a similar principle. I’m agnostic whether conventional Turing machine or quantum processing would be used, or any combination. In any case, it isn’t my problem, I suspect that future AIs will figure out the physics and invent the appropriate IT.

Processing devices operating at high speed could use a lot of energy and generate a lot of heat, and encouraging the system to lase by design would be a good way to cool it as well as powering it.

A processor using such mechanisms need not be bulky. I always assumed a yogurt pot size for my gel computer before and an air gel processor could be the same, about 100ml. That is enough to suspend a trillion particles with good line of sight for optical interconnections, and each connection could utilise up to millions of alternative wavelengths. Each wavelength could support many TDM channels and spinning the particles multiplies that up again. A UV laser clock/power source driving processors at 10^16Hz would certainly need to use high density multiplexing to make use of such a volume, with transmission distances up to 10cm (but most sub-mm) otherwise being a strongly limiting performance factor, but 10 million-fold WDM/TDM is attainable.

A trillion of these hyper-neurons using that multiplexing would act very effectively as 10 million trillion neurons, each operating at 10^16Hz processing speed. That’s quite a lot of zeros, 35 of them, and yet each hyperneuron could have connections to thousands of others in each of many physical configurations. It would be an obvious platform for supporting a large population of electronically immortal people and AIs who each want a billion replicas, and if it only occupies 100ml of space, the environmental footprint isn’t an issue.

It’s hard to know how to talk to a computer that operates like a brain, but is 10^22 times faster, but I’d suggest ‘Yes Boss’.

 

With automation driving us towards UBI, we should consider a culture tax

Regardless of party politics, most people want a future where everyone has enough to live a dignified and comfortable life. To make that possible, we need to tweak a few things.

Universal Basic Income

I suggested a long time ago that in the far future we could afford a basic income for all, without any means testing on it, so that everyone has an income at a level they can live on. It turned out I wasn’t the only one thinking that and many others since have adopted the idea too, under the now usual terms Universal Basic Income or the Citizen Wage. The idea may be old, but the figures are rarely discussed. It is harder than it sounds and being a nice idea doesn’t ensure  economic feasibility.

No means testing means very little admin is needed, saving the estimated 30% wasted on admin costs today. Then wages could go on top, so that everyone is still encouraged to work, and then all income from all sources is totalled and taxed appropriately. It is a nice idea.

The difference between figures between parties would be relatively minor so let’s ignore party politics. In today’s money, it would be great if everyone could have, say, £30k a year as a state benefit, then earn whatever they can on top. £30k is around today’s average wage. It doesn’t make you rich, but you can live on it so nobody would be poor in any sensible sense of the word. With everyone economically provided for and able to lead comfortable and dignified lives, it would be a utopia compared to today. Sadly, it can’t work with those figures yet. 65,000,000 x £30,000 = £1,950Bn . The UK economy isn’t big enough. The state only gets to control part of GDP and out of that reduced budget it also has its other costs of providing health, education, defence etc, so the amount that could be dished out to everyone on this basis is therefore a lot smaller than 30k. Even if the state were to take 75% of GDP and spend most of it on the basic income, £10k per person would be pushing it. So a couple would struggle to afford even the most basic lifestyle, and single people would really struggle. Some people would still need additional help, and that reduces the pool left to pay the basic allowance still further. Also, if the state takes 75% of GDP, only 25% is left for everything else, so salaries would be flat, reducing the incentive to work, while investment and entrepreneurial activity are starved of both resources and incentive. It simply wouldn’t work today.

Simple maths thus forces us to make compromises. Sharing resources reduces costs considerably. In a first revision, families might be given less for kids than for the adults, but what about groups of young adults sharing a big house? They may be adults but they also benefit from the same economy of shared resources. So maybe there should be a household limit, or a bedroom tax, or forms and means testing, and it mustn’t incentivize people living separately or house supply suffers. Anyway, it is already getting complicated and our original nice idea is in the bin. That’s why it is such a mess at the moment. There just isn’t enough money to make everyone comfortable without doing lots of allowances and testing and admin. We all want utopia, but we can’t afford it. Even the modest £30k-per-person utopia costs at least 3 times more than the UK can afford. Switzerland is richer per capita but even there they have rejected the idea.

However, if we can get back to the average 2.5% growth per year in real terms that used to apply pre-recession, and surely we can, it would only take 45 years to get there. That isn’t such a long time. We have hope that if we can get some better government than we have had of late, and are prepared to live with a little economic tweaking, we could achieve good quality of life for all in the second half of the century.

So I still really like the idea of a simple welfare system, providing a generous base level allowance to everyone, topped up by rewards of effort, but recognise that we in the UK will have to wait decades before we can afford to put that base level at anything like comfortable standards though other economies could afford it earlier.

Meanwhile, we need to tweak some other things to have any chance of getting there. I’ve commented often that pure capitalism would eventually lead to a machine-based economy, with the machine owners having more and more of the cash, and everyone else getting poorer, so the system will fail. Communism fails too. Thankfully much of the current drive in UBI thinking is coming from the big automation owners so it’s comforting to know that they seem to understand the alternative.

Capitalism works well when rewards are shared sensibly, it fails when wealth concentration is too high or when incentive is too low. Preserving the incentive to work and create is a mainly matter of setting tax levels well. Making sure that wealth doesn’t get concentrated too much needs a new kind of tax.

Culture tax

The solution I suggest is a culture tax. Culture in the widest sense.

When someone creates and builds a company, they don’t do so from a state of nothing. They currently take for granted all our accumulated knowledge and culture – trained workforce, access to infrastructure, machines, governance, administrative systems, markets, distribution systems and so on. They add just another tiny brick to what is already a huge and highly elaborate structure. They may invest heavily with their time and money but actually when  considered overall as part of the system their company inhabits, they only pay for a fraction of the things their company will use.

That accumulated knowledge, culture and infrastructure belongs to everyone, not just those who choose to use it. It is common land, free to use, today. Businesses might consider that this is what they pay taxes for already, but that isn’t explicit in the current system.

The big businesses that are currently avoiding paying UK taxes by paying overseas companies for intellectual property rights could be seen as trailblazing this approach. If they can understand and even justify the idea of paying another part of their company for IP or a franchise, why should they not pay the host country for its IP – access to the residents’ entire culture?

This kind of tax would provide the means needed to avoid too much concentration of wealth. A future businessman might still choose to use only software and machines instead of a human workforce to save costs, but levying taxes on use of  the cultural base that makes that possible allows a direct link between use of advanced technology and taxation. Sure, he might add a little extra insight or new knowledge, but would still have to pay the rest of society for access to its share of the cultural base, inherited from the previous generations, on which his company is based. The more he automates, the more sophisticated his use of the system, the more he cuts a human workforce out of his empire, the higher his taxation. Today a company pays for its telecoms service which pays for the network. It doesn’t pay explicitly for the true value of that network, the access to people and businesses, the common language, the business protocols, a legal system, banking, payments system, stable government, a currency, the education of the entire population that enables them to function as actual customers. The whole of society owns those, and could reasonably demand rent if the company is opting out of the old-fashioned payments mechanisms – paying fair taxes and employing people who pay taxes. Automate as much as you like, but you still must pay your share for access to the enormous value of human culture shared by us all, on which your company still totally depends.

Linking to technology use makes good sense. Future AI and robots could do a lot of work currently done by humans. A few people could own most of the productive economy. But they would be getting far more than their share of the cultural base, which belongs equally to everyone. In a village where one farmer owns all the sheep, other villagers would be right to ask for rent for their share of the commons if he wants to graze them there.

I feel confident that this extra tax would solve many of the problems associated with automation. We all equally own the country, its culture, laws, language, human knowledge (apart from current patents, trademarks etc. of course), its public infrastructure, not just businessmen. Everyone surely should have the right to be paid if someone else uses part of their share. A culture tax would provide a fair ethical basis to demand the taxes needed to pay the Universal basic Income so that all may prosper from the coming automation.

The extra culture tax would not magically make the economy bigger, though automation may well increase it a lot. The tax would ensure that wealth is fairly shared. Culture tax/UBI duality is a useful tool to be used by future governments to make it possible to keep capitalism sustainable, preventing its collapse, preserving incentive while fairly distributing reward. Without such a tax, capitalism simply may not survive.