AI is mainly a stimulative technology that will create jobs

AI has been getting a lot of bad press the last few months from doom-mongers predicting mass unemployment. Together with robotics, AI will certainly help automate a lot of jobs, but it will also create many more and will greatly increase quality of life for most people. By massively increasing the total effort available to add value to basic resources, it will increase the size of the economy and if that is reasonably well managed by governments, that will be for all our benefit. Those people who do lose their jobs and can’t find or create a new one could easily be supported by a basic income financed by economic growth. In short, unless government screws up, AI will bring huge benefits, far exceeding the problems it will bring.

Over the last 20 years, I’ve often written about the care economy, where the more advanced technology becomes, the more it allows to concentrate on those skills we consider fundamentally human – caring, interpersonal skills, direct human contact services, leadership, teaching, sport, the arts, the sorts of roles that need emphatic and emotional skills, or human experience. AI and robots can automate intellectual and physical tasks, but they won’t be human, and some tasks require the worker to be human. Also, in most careers, it is obvious that people focus less and less on those automatable tasks as they progress into the most senior roles. Many board members in big companies know little about the industry they work in compared to most of their lower paid workers, but they can do that job because being a board member is often more about relationships than intellect.

AI will nevertheless automate many tasks for many workers, and that will free up much of their time, increasing their productivity, which means we need fewer workers to do those jobs. On the other hand, Google searches that take a few seconds once took half a day of research in a library. We all do more with our time now thanks to such simple AI, and although all those half-days saved would add up to a considerable amount of saved work, and many full-time job equivalents, we don’t see massive unemployment. We’re all just doing better work. So we can’t necessarily conclude that increasing productivity will automatically mean redundancy. It might just mean that we will do even more, even better, like it has so far. Or at least, the volume of redundancy might be considerably less. New automated companies might never employ people in those roles and that will be straight competition between companies that are heavily automated and others that aren’t. Sometimes, but certainly not always, that will mean traditional companies will go out of business.

So although we can be sure that AI and robots will bring some redundancy in some sectors, I think the volume is often overestimated and often it will simply mean rapidly increasing productivity, and more prosperity.

But what about AI’s stimulative role? Jobs created by automation and AI. I believe this is what is being greatly overlooked by doom-mongers. There are three primary areas of job creation:

One is in building or programming robots, maintaining them, writing software, or teaching them skills, along with all the associated new jobs in supporting industry and infrastructure change. Many such jobs will be temporary, lasting a decade or so as machines gradually take over, but that transition period is extremely valuable and important. If anything, it will be a lengthy period of extra jobs and the biggest problem may well be filling those jobs, not widespread redundancy.

Secondly, AI and robots won’t always work direct with customers. Very often they will work via a human intermediary. A good example is in medicine. AI can make better diagnoses than a GP, and could be many times cheaper, but unless the patient is educated, and very disciplined and knowledgeable, it also needs a human with human skills to talk to a patient to make sure they put in correct information. How many times have you looked at an online medical diagnosis site and concluded you have every disease going? It is hard to be honest sometimes when you are free to interpret every possible symptom any way you want, much easier to want to be told that you have a special case of wonderful person syndrome. Having to explain to a nurse or technician what is wrong forces you to be more honest about it. They can ask you similar questions, but your answers will need to be moderated and sensible or you know they might challenge you and make you feel foolish. You will get a good diagnosis because the input data will be measured, normalized and scaled appropriately for the AI using it. When you call a call center and talk to a human, invariably they are already the front end of a massive AI system. Making that AI bigger and better won’t replace them, just mean that they can deal with your query better.

Thirdly, and I believe most importantly of all, AI and automation will remove many of the barriers that stop people being entrepreneurs. How many business ideas have you had and not bothered to implement because it was too much effort or cost or both for too uncertain a gain? 10? 100? 1000? Suppose you could just explain your idea to your home AI and it did it all for you. It checked the idea, made a model, worked out how to make it work or whether it was just a crap idea. It then explained to you what the options were and whether it would be likely to work, and how much you might earn from it, and how much you’d actually have to do personally and how much you could farm out to the cloud. Then AI checked all the costs and legal issues, did all the admin, raised the capital by explaining the idea and risks and costs to other AIs, did all the legal company setup, organised the logistics, insurance, supply chains, distribution chains, marketing, finance, personnel, ran the payroll and tax. All you’d have to do is some of the fun work that you wanted to do when you had the idea and it would find others or machines or AI to fill in the rest. In that sort of world, we’d all be entrepreneurs. I’d have a chain of tea shops and a fashion empire and a media empire and run an environmental consultancy and I’d be an artist and a designer and a composer and a genetic engineer and have a transport company and a construction empire. I don’t do any of that because I’m lazy and not at all entrepreneurial, and my ideas all ‘need work’ and the economy isn’t smooth and well run, and there are too many legal issues and regulations and it would all be boring as hell. If we automate it and make it run efficiently, and I could get as much AI assistance as I need or want at every stage, then there is nothing to stop me doing all of it. I’d create thousands of jobs, and so would many other people, and there would be more jobs than we have people to fill them, so we’d need to build even more AI and machines to fill the gaps caused by the sudden economic boom.

So why the doom? It isn’t justified. The bad news isn’t as bad as people make out, and the good news never gets a mention. Adding it together, AI will stimulate more jobs, create a bigger and a better economy, we’ll be doing far more with our lives and generally having a great time. The few people who will inevitably fall through the cracks could easily be financed by the far larger economy and the very generous welfare it can finance. We can all have the universal basic income as our safety net, but many of us will be very much wealthier and won’t need it.

 

Explaining accelerating universe expansion without dark energy

I am not the only ex-physicist that doesn’t believe in dark matter or dark energy, or multiple universes. All of these are theoretically possible interpretations of the maths, but I do not believe they are interpretations appropriate to our universe. Like the concept of the ether, I expect they will be shown to be incorrect and replaced by explanations that don’t need such concepts.

There are already explanations for accelerating expansion that don’t rely on dark energy, such as relativity: https://astronomynow.com/2015/01/05/dark-energy-explained-by-relativistic-time-dilation/ (the title is confusing since the article explains why it isn’t needed).

My theory is even simpler and probably not original, but I can’t find any references to it on the first two pages of Google so either it’s novel or so wrong that it doesn’t even warrant mentions. Anyway, here it is, make up your own mind, it doesn’t even need equations to explain it:

As galaxies get further apart, the various field fluxes reduce with the square of distance – gravitational, electromagnetic, and so must the Higgs flux. The Higgs field is what gives particles their mass. As the Higgs field declines, the mass of the particles in each galaxy must therefore drop too. If energy is to be conserved, then as mass declines, Galaxy speed must increase linearly with distance, as is the observation. QED.

Google v Facebook – which contributes most to humanity?

Please don’t take this too seriously, it’s intended as just a bit of fun. All of it is subjective and just my personal opinion of the two companies.

Google’s old motto of ‘do no evil’ has taken quite a battering over the last few years, but my overall feeling towards them remains somewhat positive overall. Facebook’s reputation has also become muddied somewhat, but I’ve never been an active user and always found it supremely irritating when I’ve visited to change privacy preferences or read a post only available there, so I guess I am less positive towards them. I only ever post to Facebook indirectly via this blog and twitter. On the other hand, both companies do a lot of good too. It is impossible to infer good or bad intent because end results arise from a combination of intent and many facets of competence such as quality of insight, planning, competence, maintenance, response to feedback and many others. So I won’t try to differentiate intent from competence and will just stick to casual amateur observation of the result. In order to facilitate score-keeping of the value of their various acts, I’ll use a scale from very harmful to very beneficial, -10 to +10.

Google (I can’t bring myself to discuss Alphabet) gave us all an enormous gift of saved time, improved productivity and better self-fulfilment by effectively replacing a day in the library with a 5 second online search. We can all do far more and live richer lives as a result. They have continued to build on that since, adding extra features and improved scope. It’s far from perfect, but it is a hell of a lot better than we had before. Score: +10

Searches give Google a huge and growing data pool covering the most intimate details of every aspect of our everyday lives. You sort of trust them not to blackmail you or trash your life, but you know they could. The fact remains that they actually haven’t. It is possible that they might be waiting for the right moment to destroy the world, but it seems unlikely. Taking all our intimate data but choosing not to end the world yet: Score +9

On the other hand, they didn’t do either of those things purely through altruism. We all pay a massive price: advertising. Advertising is like a tax. Almost every time you buy something, part of the price you pay goes to advertisers. I say almost because Futurizon has never paid a penny yet for advertising and yet we have sold lots, and I assume that many other organisations can say the same, but most do advertise, and altogether that siphons a huge amount from our economy. Google takes lots of advertising revenue, but if they didn’t take it, other advertisers would, so I can only give a smallish negative for that: Score -3

That isn’t the only cost though. We all spend very significant time getting rid of ads, wasting time by clicking on them, finding, downloading and configuring ad-blockers to stop them, re-configuring them to get entry to sites that try to stop us from using ad-blockers, and often paying per MB for unsolicited ad downloads to our mobiles. I don’t need to quantify that to give all that a score of -9.

They are still 7 in credit so they can’t moan too much.

Tax? They seem quite good at minimizing their tax contributions, while staying within the letter of the law, while also paying good lawyers to argue what the letter of the law actually says. Well, most of us try at least a bit to avoid paying taxes we don’t have to pay. Google claims to be doing us all a huge favor by casting light on the gaping holes in international tax law that let them do it, much like a mugger nicely shows you the consequences of inadequate police coverage by enthusiastically mugging you. Noting the huge economic problems caused across the world by global corporates paying far less tax than would seem reasonable to the average small-business-owner, I can’t honestly see how this could live comfortably with their do-no evil mantra. Score: -8

On the other hand, if they paid all that tax, we all know governments would cheerfully waste most of it. Instead, Google chooses to do some interesting things with it. They gave us Google Earth, which at least morally cancels out their ‘accidental’ uploading of everyone’s wireless data as their street-view cars went past.They have developed self-driving cars. They have bought and helped develop Deep-mind and their quantum computer. They have done quite a bit for renewable energy. They have spent some on high altitude communications planes supposedly to bring internet to the rural parts of the developing world. When I were a lad, I wanted to be a rich bastard so I could do all that. Now, I watch as the wealthy owners of these big companies do it instead. I am fairly happy with that. I get the results and didn’t have to make the effort. We get less tax, but at least we get some nice toys. Almost cancels. Score +6

They are trying to use their AI to analyse massive data pools of medical records to improve medicine. Score +2

They are also building their databases more while doing that but we don’t yet see the downside. We have to take what they are doing on trust until evidence shows otherwise.

Google has tried and failed at many things that were going to change the world and didn’t, but at least they tried. Most of us don’t even try. Score +2

Oh yes, they bought YouTube, so I should factor that in. Mostly harmless and can be fun. Score: +2

Almost forgot Gmail too. Score +3

I’m done. Total Google contribution to humanity: +14

Well done! Could do even better.

I’ve almost certainly overlooked some big pluses and minuses, but I’ll leave it here for now.

Now Facebook.

It’s obviously a good social network site if you want that sort of thing. It lets people keep in touch with each other, find old friends and make new ones. It lets others advertise their products and services, and others to find or spread news. That’s all well and good and even if I and many other people don’t want it, many others do, so it deserves a good score, even if it isn’t as fantastic as Google’s search, that almost everyone uses, all the time. Score +5

Connected, but separate from simply keeping in touch, is the enormous pleasure value people presumably get from socializing. Not me personally, but ‘people’. Score +8

On the downside: Quite a lot of problems result from people, especially teens, spending too much time on Facebook. I won’t reproduce the results of all the proper academic  studies here, but we’ve all seen various negative reports: people get lower grades in their exams, people get bullied, people become socially competitive – boasting about their successes while other people feel insecure or depressed when others seem to be doing better, or are prettier, or have more friends. Keeping in touch is good, but cutting bits off others’ egos to build your own isn’t. It is hard not to conclude that the negative uses of keeping in touch outweigh the positive ones. Long-lived bad-feelings outweigh short-lived ego-boosts. Score: -8

Within a few years of birth, Facebook evolved from a keeping-in-touch platform to a general purpose mini-web. Many people were using Facebook to do almost everything that others would do on the entire web. Being in a broom cupboard is fine for 5 minutes if you’re playing hide and seek, but it is not desirable as a permanent state. Still, it is optional, so isn’t that bad per se: Score: -3

In the last 2 or 3 years, it has evolved further, albeit probably unintentionally, to become a political bubble, as has become very obvious in Brexit and the US Presidential Election, though it was already apparent well before those. Facebook may not have caused the increasing divide we are seeing between left and right, across the whole of the West, but it amplifies it. Again, I am not implying any intent, just observing the result. Most people follow people and media that echoes their own value judgments. They prefer resonance to dissonance. They prefer to have their views reaffirmed than to be disputed. When people find a comfortable bubble where they feel they belong, and stay there, it is easy for tribalism to take root and flourish, with demonization of the other not far behind. We are now seeing that in our bathtub society, with two extremes and a rapidly shallowing in-between that was not long ago the vast majority. Facebook didn’t create human nature; rather, it is a victim of it, but nonetheless it provides a near-monopoly social network that facilitates such political bubbles and their isolation while doing far too little to encourage integration in spite of its plentiful resources. Dangerous and Not Good. Score -10

On building databases of details of our innermost lives, managing not to use the data to destroy our lives but instead only using it to sell ads, they compare with Google. I’ll score that the same total for the same reasons: Net Score -3

Tax? Quantities are different, but eagerness to avoid tax seems similar to Google. Principles matter. So same score: -8

Assorted messaging qualifies as additional to the pure social networking side I think so I’ll generously give them an extra bit for that: Score +2

They occasionally do good things with it like Google though. They also are developing a high altitude internet, and are playing with space exploration. Tiny bit of AI stuff, but not much else has crossed my consciousness. I think it is far less than Google but still positive, so I’ll score: +3

I honestly can’t think of any other significant contributions from Facebook to make the balance more positive, and I tried. I think they want to make a positive contribution, but are too focused on income to tackle the social negatives properly.

Total Facebook contribution to humanity: -14.

Oh dear! Must do better.

Conclusion: We’d be a lot worse off without Google. Even with their faults, they still make a great contribution to humankind. Maybe not quite a ‘do no evil’ rating, but certainly they qualify for ‘do net good’. On the other hand, sadly, I have to say that my analysis suggests we’d be a lot better off without Facebook. As much better off without them as we benefit by having Google.

If I have left something major out, good or bad, for either company please feel free to add your comments. I have deliberately left out their backing of their own political leanings and biases because whether you think they are good or bad depends where you are coming from. They’d only score about +/-3 anyway, which isn’t a game changer.

 

 

Future sex, gender and relationships: how close can you get?

Using robots for gender play

Using robots for gender play

I recently gave a public talk at the British Academy about future sex, gender, and relationship, asking the question “How close can you get?”, considering particularly the impact of robots. The above slide is an example. People will one day (between 2050 and 2065 depending on their budget) be able to use an android body as their own or even swap bodies with another person. Some will do so to be young again, many will do so to swap gender. Lots will do both. I often enjoy playing as a woman in computer games, so why not ‘come back’ and live all over again as a woman for real? Except I’ll be 90 in 2050.

The British Academy kindly uploaded the audio track from my talk at

If you want to see the full presentation, here is the PowerPoint file as a pdf:

sex-and-robots-british-academy

I guess it is theoretically possible to listen to the audio while reading the presentation. Most of the slides are fairly self-explanatory anyway.

Needless to say, the copyright of the presentation belongs to me, so please don’t reproduce it without permission.

Enjoy.

Christmas in 2040

I am cheating with this post, since I did a newspaper interview that writes up some of my ideas and will save time rewriting it all. Here’s a link:

https://www.thesun.co.uk/living/2454633/dinner-cooked-by-robots-no-wrapping-paper-and-video-make-up-for-the-office-party-this-is-what-christmas-will-look-like-in-2040-according-to-futurologist-dr-ian-pearson/

I hope you all have a wonderful Christmas.

Get all of my current e-books free, today only

This offer is now over. Sorry if you missed it.

As an early Christmas present, I have made all of my books free just for today on Amazon. The links here are for amazon.co.uk, but the book reference is the same on other branches so just edit the .co.uk to .com or whatever.

You Tomorrow and Society Tomorrow were almost entirely made by adding some of my blogs, tidying up and filling a few gaps.

https://www.amazon.co.uk/You-Tomorrow-Ian-Pearson-ebook/dp/B00G8DLB24

https://www.amazon.co.uk/Society-Tomorrow-Growing-Century-Britain-ebook/dp/B01HJY7RHI

Total Sustainability takes a system level view of sustainability and contradicts a lot of environmentalist dogma.

https://www.amazon.co.uk/Total-Sustainability-Ian-Pearson-ebook/dp/B00FWMW194

Space Anchor is my only Sci-fi novel to date, and features the first ever furry space ship in sci-fi, a gender-fluid AI, and its heroes Carbon Girl and Carbon Man have an almost entirely carbon-based itinerary.

https://www.amazon.co.uk/Space-Anchor-Ian-Pearson-ebook/dp/B00E9X02IE

Enjoy reading. Next year I hope to finish my book on future fashion.

 

Fluorescent microsphere mist displays

A few 3D mist displays have been demonstrated over the last decade. I’ve seen a couple at trade shows and have been impressed. To date, they use mists or curtains of tiny water droplets to make a 3D space onto which to project an image, so you get a walk-through 3D life-sized display. Like this:

http://wonderfulengineering.com/leia-display-system-uses-a-screen-made-of-water-mist-to-display-3d-projections/

or check out: http://ixfocus.com/top-10-best-3d-water-projections-ever/

Two years ago, I suggested using a forehead-mounted mist projector:

https://timeguide.wordpress.com/2014/11/03/forehead-3d-mist-projector/

so you could have a 3D image made right in front of you anywhere.

This week, a holographic display has been doing the rounds on Twitter, called Gatebox:

https://www.geek.com/tech/gatebox-wants-to-be-your-personal-holographic-companion-1682967/

It looks OK but mist displays might be better solution for everyday use because they can be made a lot bigger more cheaply. However, nobody really wants water mist causing electrical problems in their PCs or making their notebook paper soggy. You can use smoke as a mist substitute but then you have a cloud of smoke around you. So…

Suppose instead of using water droplets and walking around veiled in fog or smoke or accompanied by electrical crackling and dead PCs, that the mist was not made of water droplets but tiny dry and obviously non-toxic particles such as fluorescent micro-spheres that are invisible to the naked eye and transparent to visible light so you can’t see the mist at all, and it won’t make stuff damp. Instead of projecting visible light, the particles are made of fluorescent material, so that they are illuminated by a UV projector and fluoresce with the right colour to make the visible display. There are plenty of fluorescent materials that could be made into tiny particles, even nano-particles, and made into an invisible mist that produces a bright and high-resolution display. Even if non-toxic is too big an ask, or the fluorescent material is too expensive to waste, a large box that keeps them contained and recycles them for the next display could still be bigger, better, brighter and cheaper than a large holographic display.

Remember, you saw it here first. My 101st invention of 2016.

Chat-bots will help reduce loneliness, a bit

Amazon is really pushing its Echo and Dot devices at the moment and some other companies also use Alexa in their own devices. They are starting to gain avatar front ends too. Microsoft has their Cortana transforming into Zo, Apple has Siri’s future under wraps for now. Maybe we’ll see Siri in a Sari soon, who knows. Thanks to rapidly developing AI, chatbots and other bots have also made big strides in recent years, so it’s obvious that the two can easily be combined. The new voice control interfaces could become chatbots to offer a degree of companionship. Obviously that isn’t as good as chatting to real people, but many, very many people don’t have that choice. Loneliness is one of the biggest problems of our time. Sometimes people talk to themselves or to their pet cat, and chatting to a bot would at least get a real response some of the time. It goes further than simple interaction though.

I’m not trying to understate the magnitude of the loneliness problem, and it can’t solve it completely of course, but I think it will be a benefit to at least some lonely people in a few ways. Simply having someone to chat to will already be of some help. People will form emotional relationships with bots that they talk to a lot, especially once they have a visual front end such as an avatar. It will help some to develop and practice social skills if that is their problem, and for many others who feel left out of local activity, it might offer them real-time advice on what is on locally in the next few days that might appeal to them, based on their conversations. Talking through problems with a bot can also help almost as much as doing so with a human. In ancient times when I was a programmer, I’d often solve a bug by trying to explain how my program worked, and in doing so i would see the bug myself. Explaining it to a teddy bear would have been just as effective, the chat was just a vehicle for checking through the logic from a new angle. The same might apply to interactive conversation with a bot. Sometimes lonely people can talk too much about problems when they finally meet people, and that can act as a deterrent to future encounters, so that barrier would also be reduced. All in all, having a bot might make lonely people more able to get and sustain good quality social interactions with real people, and make friends.

Another benefit that has nothing to do with loneliness is that giving a computer voice instructions forces people to think clearly and phrase their requests correctly, just like writing a short computer program. In a society where so many people don’t seem to think very clearly or even if they can, often can’t express what they want clearly, this will give some much needed training.

Chatbots could also offer challenges to people’s thinking, even to help counter extremism. If people make comments that go against acceptable social attitudes or against known facts, a bot could present the alternative viewpoint, probably more patiently than another human who finds such viewpoints frustrating. I’d hate to see this as a means to police political correctness, though it might well be used in such a way by some providers, but it could improve people’s lack of understanding of even the most basic science, technology, culture or even politics, so has educational value. Even if it doesn’t convert people, it might at least help them to understand their own views more clearly and be better practiced at communicating their arguments.

Chat bots could make a significant contribution to society. They are just machines, but those machines are tools for other people and society as a whole to help more effectively.

 

25 predictions for 2017

2017-predictions

AI presents a new route to attack corporate value

As AI increases in corporate, social, economic and political importance, it is becoming a big target for activists and I think there are too many vulnerabilities. I think we should be seeing a lot more articles than we are about what developers are doing to guard against deliberate misdirection or corruption, and already far too much enthusiasm for make AI open source and thereby giving mischief-makers the means to identify weaknesses.

I’ve written hundreds of times about AI and believe it will be a benefit to humanity if we develop it carefully. Current AI systems are not vulnerable to the terminator scenario, so we don’t have to worry about that happening yet. AI can’t yet go rogue and decide to wipe out humans by itself, though future AI could so we’ll soon need to take care with every step.

AI can be used in multiple ways by humans to attack systems.

First and most obvious, it can be used to enhance malware such as trojans or viruses, or to optimize denial of service attacks. AI enhanced security systems already battle against adaptive malware and AI can probe systems in complex ways to find vulnerabilities that would take longer to discover via manual inspection. As well as AI attacking operating systems, it can also attack AI by providing inputs that bias its learning and decision-making, giving AI ‘fake news’ to use current terminology. We don’t know the full extent of secret military AI.

Computer malware will grow in scope to address AI systems to undermine corporate value or political campaigns.

A new route to attacking corporate AI, and hence the value in that company that relates in some way to it is already starting to appear though. As companies such as Google try out AI-driven cars or others try out pavement/sidewalk delivery drones, so mischievous people are already developing devious ways to misdirect or confuse them. Kids will soon have such activity as hobbies. Deliberate deception of AI is much easier when people know how they work, and although it’s nice for AI companies to put their AI stuff out there into the open source markets for others to use to build theirs, that does rather steer future systems towards a mono-culture of vulnerability types. A trick that works against one future AI in one industry might well be adaptable to another use in another industry with a little devious imagination. Let’s take an example.

If someone builds a robot to deliberately step in front of a self-driving car every time it starts moving again, that might bring traffic to a halt, but police could quickly confiscate the robot, and they are expensive, a strong deterrent even if the pranksters are hiding and can’t be found. Cardboard cutouts might be cheaper though, even ones with hinged arms to look a little more lifelike. A social media orchestrated campaign against a company using such cars might involve thousands of people across a country or city deliberately waiting until the worst time to step out into a road when one of their vehicles comes along, thereby creating a sort of denial of service attack with that company seen as the cause of massive inconvenience for everyone. Corporate value would obviously suffer, and it might not always be very easy to circumvent such campaigns.

Similarly, the wheeled delivery drones we’ve been told to expect delivering packages any time soon will also have cameras to allow them to avoid bumping into objects or little old ladies or other people, or cats or dogs or cardboard cutouts or carefully crafted miniature tank traps or diversions or small roadblocks that people and pets can easily step over but drones can’t, that the local kids have built from a few twigs or cardboard from a design that has become viral that day. A few campaigns like that with the cold pizzas or missing packages that result could severely damage corporate value.

AI behind websites might also be similarly defeated. An early experiment in making a Twitter chat-bot that learns how to tweet by itself was quickly encouraged by mischief-makers to start tweeting offensively. If people have some idea how an AI is making its decisions, they will attempt to corrupt or distort it to their own ends. If it is heavily reliant on open source AI, then many of its decision processes will be known well enough for activists to develop appropriate corruption tactics. It’s not to early to predict that the proposed AI-based attempts by Facebook and Twitter to identify and defeat ‘fake news’ will fall right into the hands of people already working out how to use them to smear opposition campaigns with such labels.

It will be a sort of arms race of course, but I don’t think we’re seeing enough about this in the media. There is a great deal of hype about the various AI capabilities, a lot of doom-mongering about job cuts (and a lot of reasonable warnings about job cuts too) but very little about the fight back against AI systems by attacking them on their own ground using their own weaknesses.

That looks to me awfully like there isn’t enough awareness of how easily they can be defeated by deliberate mischief or activism, and I expect to see some red faces and corporate account damage as a result.

PS

This article appeared yesterday that also talks about the bias I mentioned: https://techcrunch.com/2016/12/10/5-unexpected-sources-of-bias-in-artificial-intelligence/

Since I wrote this blog, I was asked via Linked-In to clarify why I said that Open Source AI systems would have more security risk. Here is my response:

I wasn’t intending to heap fuel on a dying debate (though since current debate looks the same as in early 1990s it is dying slowly). I like and use open source too. I should have explained my reasoning better to facilitate open source checking: In regular (algorithmic) code, programming error rate should be similar so increasing the number of people checking should cancel out the risk from more contributors so there should be no a priori difference between open and closed. However:

In deep learning, obscurity reappears via neural net weightings being less intuitive to humans. That provides a tempting hiding place.

AI foundations are vulnerable to group-think, where team members share similar world models. These prejudices will affect the nature of OS and CS code and result in AI with inherent and subtle judgment biases which will be less easy to spot than bugs and be more visible to people with alternative world models. Those people are more likely to exist in an OS pool than a CS pool and more likely to be opponents so not share their results.

Deep learning may show the equivalent of political (or masculine and feminine). As well as encouraging group-think, that also distorts the distribution of biases and therefore the cancelling out of errors can no longer be assumed.

Human factors in defeating security often work better than exploiting software bugs. Some of the deep learning AI is designed to mimic humans as well as possible in thinking and in interfacing. I suspect that might also make them more vulnerable to meta-human-factor attacks. Again, exposure to different and diverse cultures will show a non-uniform distribution of error/bias spotting/disclosure/exploitation.

Deep learning will become harder for humans to understand as it develops and becomes more machine dependent. That will amplify the above weaknesses. Think of optical illusions that greatly distort human perception and think of similar in advanced AI deep learning. Errors or biases that are discovered will become more valuable to an opponent since they are less likely to be spotted by others, increasing their black market exploitation risk.

I have not been a programmer for over 20 years and am no security expert so my reasoning may be defective, but at least now you know what my reasoning was and can therefore spot errors in it.