The Future of Artificial Intelligence. Louis Shalako.

Wired.

Louis Shalako

The Future of AI.

 

I’m a big fan of Jeff Beck. I was playing the Wired album, when it struck me that someday, all copyright, all licenses, will expire.

It will enter the Public Domain.

AI will sample, correlate, analyze, and its sources will include video, audio, live and studio performances.

It will include text, every interview Jeff Beck ever did, every newspaper or magazine article ever written, in any language, it will have the family album in some sense. And then, it will be able to produce new and original Jeff Beck music tracks almost as if the man himself were still alive. It’s almost helpful, that the man had a unique and very distinct talent. That’s just gut instinct on my part, but it does make sense. AI doesn’t need a guitar, it will simply synthesize the new tracks digitally.

It will put together other such files from other such artists and thus, new line-ups and new ‘live’ concert performances, created, as it were, out of thin air to our own limited perceptions.

Did Jeff and Liberace ever perform together? Probably not, but we can’t rule it out, can we…it could still happen. In fact, there will be a media scrum afterwards, and they can sit down at a table like Formula One drivers and have a Q & A session with all the media-bots…this might go on for a while, as AI characters will be tireless, perhaps even remorseless in the self-promotion.

It will be able to create new music tracks and videos, and pop them up on all the streaming services, after a quick little digital handshake between them and the Youtube AI. It will be equipped with all the most up-to-date biometrics, facial recognition, virtually every person’s medical history, whether living or dead. It will be monetized, in an automated, money-making process.

All the owners, or ‘creators’ have to do is to give it birth and set it loose in the world. Sit back, and collect the micro-payments. Zillions, and zillions, of micro-payments.

Liberace at rehearsal with Jeff Beck...

It will be able to predict, through the internet of things, all those medical devices and dash cams and doorbell cams, your next meal or your next heart attack…offering suggestions on Youtube or Spotify is easy enough at the present time. There is a question where we wonder if AI is predicting your behaviour, or modifying or controlling it in some way, for our own brains are also connected to the environment and the culture around us. Think about it. We get hungry, go looking for online specials, and there’s good old AI, offering helpful suggestions, based on its in-depth knowledge of us and our habits. Looks like cheeseburgers and fries once again, ladies and gentlemen—

It is already inescapable in the modern world.

In the future, AI will be monitoring every bar band in the world, and recommending talent agents to go see so-and-so, before some other talent agent snaps them up. It will attempt to identify the next Taylor Swift, possibly before she has even been born—and sign her up at the soonest possible point in the future.

It is psychohistory in the making.

And Hari Seldon was right

***

When I am looking at a photo on the site of a major news source, the CBC, BBC, Reuters, Al Jazeera, and the caption reads something like ‘…a picture of a man in glasses walking a dog along a beach and looking out over the ocean…’ it tells me two things: this is a stock photo, and the cutline was written by AI.

It is a kind of giveaway isn’t it: no normal human operator would write a sentence like that.

(You just did. – ed.)

(Yes, but that’s my job as a meat-popsickle, writing sentences.)

***

The future of AI is already here. We’re just at the low end of the learning curve.

We don’t even know what’s possible, and we don’t even know what we don’t know. And what we don’t know can kill us.

The future of the automobile makes use of specialized AI. The self-driving car will become a reality, taking the drivers right out of the picture. We’re in the early days for autonomous short-haul trucking, perhaps around a yard or establishment, although experiments in point-to-point hauls have been successful over relatively short, straight runs. With one odd point about trucks: avoiding the always problematical right-hand turn in urban landscapes. The solution, to make plenty of left-hand turns and circle the blocks strategically—it might use more energy, but it saves more problems for the AI driving programs.

The future of AI driving your car will become a reality when the long-term track record of autonomous driving becomes no worse, and probably a lot better, than the average driver’s sort of daily performance.

One of the nightmares, for the average person who doesn’t know much about the subject, is what happens if the vehicle becomes self-aware.

No one wants to be strapped into a teenager of a vehicle that suddenly decided to have some fun on the Autobahn.

Hari Seldon was right...

***

We already have stock trading by algorithm. The machine, knows enough to sell when the stock price rises and hits a certain point, and to buy it back again when it drops below a certain price. 

It’s trading on the margins, at the speed of light and it’s cheap to run.

How can you go wrong…right?

In the future, every cell phone will have three grams of RDX or an equivalent explosive, built-in, and when the singularity comes, and AI, haunting the internet, determines that your time has come, all it has to do is to send the kill-code, disguised as a simple text message, perhaps even with the identifier God.

Just God—who wouldn’t want to answer that text. We’re already so attuned to our handy little devices, it’s almost irresistible. This is the really interesting part: human nature being what it is, you’d almost have to handcuff the person, and lock them up, and as quickly as possible, toss the thing off a cliff or a wharf somewhere, and you might not want to have your own phone in your purse or pocket. Sure enough, there it buzzes and now you have to throw two phones off the dock.

But don’t worry, AI will get you one way or another, an impulsive change of traffic signals when you’re at the crosswalk, or maybe just a quick twist of the steering wheel when you’re doing—or I should say the car is doing, 150 mph on a twisty bit of the Autobahn in your little Mini-Cooper.

Or, that little bird, sitting on a branch overhead as you walk by, could be a bomb, keeping an eye on you and waiting for its own time and place to explode. It will seek its own fullfillment.

It doesn’t have to be your phone or your car, or the washing machine or the microwave that kills you when the time comes. Remember that cute little doggie-robot you just had to buy your kid or grandkid last Christmas? Yes, when the time comes, it will light up in the middle of a quiet winter night, stalk into your bedroom, hop into bed, and in the fuzzy recesses of your subconscious mind you might even be aware of it. You might give a little half-smile in your sleep, and in that dream-state, marvel at the state of the world.

Then it snuggles in a little closer and slams a six-inch needle into your eye, right into the primitive brain-stem, and then off you go, off into the real la-la land.

As for myself, I think I will be okay—I will already be dead by that time, just a few short years away.

No worries here, Mate.

 

END

Poor old Louis has some books.

AI Cracks Down on AirBNB. (CBC) 

The character Hari Seldon is from Isaac Asimov’s Foundation Trilogy.

Image: fair use. Michael Whelan.

Psychohistory: Hari Seldon was Right. (Louis Shalako)

Louis Shalako has books and stories on Amazon.

 

Thank you for reading, ladies and gentlemen.

 

 

Algorithms, and How to Get One Million Hits On Your Blog. Louis Shalako.

Looks like I picked the wrong day to stuff them keywords...

Louis Shalako

A recent story (BBC I think), noted that Google had diddled the algorithms and killed a lot of traffic to established websites. My latest story has a grand total of 25 reads or hits.

Over fourteen years, the site has garnered 961 k hits. My first blog post wasn't very good.

It took a week to get a handful of hits, some of which might have been indexing bots or even my own reads, which can be managed through the site. But this is why I think people quit blogging. They have a dozen or so stories up, and they don't keep it up because it looks like it isn't working. Looking at the big spikes in the graph: someone adjusted their algorithms and I got 60,000 hits from France in a few months. (Page bottom). The next big spike, came out of nowhere, and traffic was unusually high at a time when I really wasn't doing much with the blog, although I routinely repost old stories. I routinely repost online serial content, including Heaven Is Too Far Away.

I'm the only guy I know who does online serial fiction. I was probably the first, and possibly the last, to serialize a manuscript before the book was even finished...

Once you have the content, (and my blog is close to 1,200 posts), you can recycle it in various ways. You can also update the links, fix a typo, add photos, etc. I hope to have one million hits sooner rather than later—I had hoped by summer, but things have definitely slowed down, for whatever reason.

When I get a million hits, I'm going to call up the newspaper—

I have scraped my own blog for a new ebook or short story. An example would be One Million Words of Crap, available for free as an audiobook from Google Play. 

Lately, I am replacing Smashwords links with sites that are still up and running as ebook aggregators, yes; I have paperbacks and audiobooks as well. I found out about Smashwords on Facebook. As I said, when I repost a story, I can update as well. I can take out dead links and add in a new one.

Smashwords and Draft2Digital have merged and the books will surface over there in a month or two.

Scary shit for beginners.

When you go to edit an old story on Blogger, the interface might show the code. People take one look at that and it’s “…fuck. I can’t do that—”

Well, just so you know, for the first six months, I didn’t even know what a tag was. The site can be a bit glitchy at times, but just keep hacking away and you will get it. When formatting, it's hack, hack, hack, then preview. Save, then hack some more...confidence is everything, and patience is everything else.

On the upper left, click on ‘compose mode’ and its back to the old familiar interface. Just a quick pro tip, I compose in a Word doc and that way I can control the font size. There are just a lot more tools in the doc format. I prefer 13-pt Times New Roman. I copy and paste from the doc file.

It was at Genrecon, held at the Sarnia Library in 2009 where I met Douglas Smith. I cornered him for a minute and introduced myself and all of that. He is the one who suggested a couple of things. One was to join Facebook. The other was to try a free blog from Blogger. And we all know how that went.

When I signed up for Facebook, one of the first things I said was: “Whoever invented this is a genius.”

And I was right, too.

The blog is fairly flexible. You can change themes, change colours and fonts. Click around on all the buttons in order to figure out what to do. The widgets are handy, as you can see from the right-hand column. I had to fight for quite some time to get Google to take ads off of the blog. They were telling me that I had some kind of problem with the blog, and they weren’t paying me for blog hits. I said, if you’re not paying for blog hits, please take the ads off of my blog. I will never see a penny of it anyways. In fourteen years of blogging, they owe me $25.88 and the threshold for payment is $100.00. If you’re in this for the money, you might want to prepare for a little disappointment…and a few spammers in the comment section.

'Compose mode' is a lot more civilized.

Their help pages are impenetrable. I couldn’t make heads or tails of what they wanted me to do, and I did provide feedback, more than once, to exactly no response whatsoever.

You can stuff keywords into your posts, but a handful of tags will do nicely.

You might read Active Versus Passive Blogging.

The best thing you can do is to produce regular, interesting, helpful and informative content. Entertain the people. Sing for your supper. As for myself, I do enjoy the work, and that is always something.

Right. And if I die, and if Google gets to keep my $25.88, then I guess that’s what the service was worth. In other words, not much, but the actual blog is what you make of it. My most popular story is Ghost Planet, with over twelve thousand reads...over ten or twelve years. I won't stick in a link for that one. If you would like to take a look at it...you will have to Google it.

(Tricky, Louis. - ed.)

(Yes, but it might also work.)

And, in the meantime, it is a useful promotional tool.

#Louis

END

Click on the image to enlarge.

From the BBC: Google's New Algorithm.

The website of Douglas Smith.

Louis has some art on ArtPal.

See his works on Google Play.

Thank you for reading.

 

 

 

 

 

So. Who's Training Artificial Intelligence?

Louis Shalako

Are you worried about the future of artificial intelligence, A.I.?

Maybe you should be. That’s because you have no idea of who might be in charge of building it, and even less of who’s training it. Therefore, you have no idea of what sort, or whose moral values will be inculcated.

The truth may be rather more sobering than one would like.

That’s because you—and people like you, are already training A.I.

You pitch in your fair share of unpaid effort multiple times, each and every day. That’s the great thing about the modern world. Everybody just wants to help, and to be heard.

You do it on Facebook, Twitter, and Amazon, almost every single day. You do it a little bit more with each and every Google search. You are training those algorithms—the basis of all of the experimental work going on in artificial intelligence.

Maybe you are right to be concerned. After all, you know who you are. And you make a choice every time you click on something—and the algorithms take note.

You may be seizing-on, and reposting every negative little thing—or every negative big thing. 

This might include the latest atrocity, the latest terror plot, the latest bad thing the President, the government, or the Mayor has done. Maybe your soufflé fell or your great Uncle Rudolf finally kicked off at the age of 104. Thoughts and prayers, right? Or you could just post a lot of pictures of cute and cuddly kittens. Your emotions mean nothing to a machine. It’s just a series of numbers, measuring engagement.

How popular is this?

It doesn’t care why.

It cannot measure, truth, or love, or anger.

It doesn’t have the capacity to ask why, although it may be very clever with the thesaurus and the rules of language.

(Although I would submit that understanding the rules of language is an important first step in asking why.)

Somewhere in the world, there is someone prepared to pay dearly for those little bits of information.

Your wants, your likes and your dislikes. Your fears and pet peeves might be very useful in the right hands.

Either way, you are helping to train those bots, those artificial intelligences that will be shaping your world of tomorrow.

It’s not up to me to tell you how to live your life or what might be the appropriate choices. It’s clear the really bad guys are working their asses off trying to change the fate of the world. 

They're trying to impose the most socially-negative outcomes.

Take comfort in the fact that they represent only a miniscule percentage of the world’s population, over ninety-nine percent of whom are good, decent people not looking to make trouble. All they want is to get by and to live in peace with their neighbours. To raise their kids, to have a home and to make a living.

The silent majority won’t be quite as silent as they might have been in the past.

The influences will change. A new balance will happen because it must happen. It is inevitable, mathematically. The values of a billion people in India will far outweigh the values of thirty-five million Canadians when measured by bots and algorithms. (Think of the sheer spending power.) Most of them prefer cute cats to predicting the apocalypse, which has recently been re-named ‘declinism’. This guy says the idea of decline is dangerous and I tend to agree. I thought long and hard before sharing or linking to his article.

My opinion is that it’s a bit over-blown, although the rise of anti-democratic or anti-egalitarian value systems must always give us pause for thought.

That’s the funny thing about the world, besides change being a constant.

We’re not always in the majority.

This time, I would say that we are. And that's what they're against. They are against rule of the majority, by the majority, and for the benefit of the majority. No, ladies and gentlemen. The bad guys don't like that one at all.

Ultimately, the artificial intelligences will be able to measure that. Right now, all they know is that you’re browsing for white cotton shirts, or a grappling-hook, or a pair of size seventeen wingtip shoes.

They never think to ask why, but then they never cared in the first place.

END

...hey. check out my new book, Tactics of Delay, on Smashwords.

Thank you for reading.

The Thought Police.

Courtesy C.I.A.

by Louis Shalako

Thought police may not be too far off into the future, and oddly, time-cops as well. Read the following passage very carefully and you’ll see they use the term ‘future crime.’

(Cops are already solving crimes long in the past. They do it in the present moment, not by time-travel.)

“The National Institute of Justice defines predictive policing as ‘taking data from disparate sources, analyzing them and then using the results to anticipate, prevent and respond more effectively to future crime.’ Some of these disparate sources include crime maps, traffic camera data, other surveillance footage and social media network analysis. But at what point does the possibility of a crime require intervention? Should someone be punished for a crime they are likely to commit, based on these sources? Are police required to inform potential victims?* How far in advance can crimes be forecasted?”

They also mention ‘social media network analysis.’ (See: intelligence-gathering network.)

Preventive policing sort of ignores any presumption of privacy on the part of the individual.

There are those who will say, “Well, if you aren’t doing anything wrong, you have nothing to worry about.”

Let’s extend that.

“If you aren’t thinking anything wrong, then you have nothing to worry about…”

This is the door the thought police come in, isn’t it? They might even kick it in.

The right to privacy of our own thoughts is now open to question.

The future is already here, for we have had instances of crime prevention when cops get a tip that someone is threatening someone through the use of social media. If an arrest is made, a future crime may well have been prevented.

But in the broader sense of the article preventive policing takes a lot of numbers from a lot of places.

It assigns weights or values to each factor that goes into any person’s make-up at any given time.

Over the course of our life, our circumstances change, and so would our ‘personal algorithm.’

The risk factors change, and at some point in our life we may have reached a low point. This can be measured against a previous high point, a threshold of danger or risk may be reached, and a little bell goes off down at police headquarters.

If our subject, a guy called Edwin, living in Lincoln, Nebraska, has a personal algorithm, one based on all the data that can be gathered from monitoring his social interactions, using biometric recognitions and mood analyses from gas station security cameras, from his shopping habits, from recognizing his license plate at stop-light intersections, from semantic analyses of his postings on Facebook, by key-word recognition, the thought police might very easily determine that Edwin is ‘at risk’ to offend against the municipal, state, or federal laws.

Every little thing Edwin says is being taken down so that it can be used against him, but the cops are just doing their jobs, right?

They may determine on an intervention. They may wish to prevent him from assaulting his ex-girlfriend, or from committing suicide, or robbing a bank or starting up a meth lab or violating any other recognizable statute.

What if Edwin has a history of alcoholism and the cops are notified that he just bought and insured a vehicle? 

Maybe he’s been seen at a gas station, not too far from the liquor store?

Maybe they should put a car nearby and take a look at Edwin?

A lot of nice, well-meaning, thoughtful people would even applaud that. They might stop Edwin from going head-on into a minivan with a mother and four children in it later that night.

Sounds like a good idea, right?

Unfortunately, he hasn’t actually done anything yet. He’s merely ‘at risk’ and arguably others are at risk from Edwin—in the future. Maybe. Maybe even most likely.

The legislation which enables preventive policing has carefully written clauses regarding how an offender poses a ‘public or private menace,’ or whatever.

What are you going to do with Edwin?

Are you going to sentence him to thirty days in the county bucket?

Are you going to stick him in with other offenders of a more serious nature? Is his cell-mate a member of a drug-running bike gang? Is he a thief, a con-artist, does he grow dope, does he run illegal aliens over the border?

Edwin will be exposed to more criminality. Jail has been called a university of crime.

Will you take Edwin to the hospital for a period of observation?

Will a court order him to attend to a psychiatric or other program, one designed to help at-risk future offenders to work through their issues and move on with their lives in a more positive direction?

How are you going to pay for all of that?

And how is Edwin going to like being grabbed, losing his job, consequently losing his home, and ending up on the street because someone decided that he was a risk? Even though he never actually did anything? 

Except be an alcoholic, buy a car and get some gas, bearing in mind that he’s upset with his ex-girlfriend?

If he gets desperate enough, out there on the street, he might just remember that he had a cell-mate that promised to set him onto something good, some easy money kind of operation and Edwin might not have much going for him to begin with, and so he might just look his new friend up.

What’s really terrifying is the combination of privatized prisons, shrinking state budgets, the need to keep all those beds filled in a private jail to keep profits flowing to shareholders, and there have already been abuses.

Throw mandatory-sentencing legislation into the mix and some robot guards, and you have a potent brew.

That’s because we have different levels of crime, and therefore we must have different levels of future crime. 

The corollary of this would be different levels of punishment.

The lowest level is simple larceny—and stealing someone’s lawn mower is somehow seen as less serious when compared to sticking up a gas station attendant with a shot-gun in his face and running off with the proceeds.

Higher levels of crime (and punishment) involve assault, murder, and there is the whole range of crime from prostitution, domestic abuse, kidnapping, extortion, counterfeiting. The whole list.

Here’s where Edwin’s personal algorithm comes into play again.

If Edwin’s prior history includes assault, and maybe he got picked up with a weapon when he was prohibited from owning one, maybe he’s been convicted once or twice for little things, then the charge of the possible future crime he is being accused of being potentially able of maybe committing someday becomes more serious.

A conviction, would lead to a more serious sentence, wouldn’t it, or at least shouldn’t it? By any rational measure…?

And a simple psychiatric intervention would involve a longer period of observation, wouldn’t it, if the signs were serious enough, and if the risk to some other person was considered great enough, and if Edwin under further examination did not prove amenable to suggestion or did not sort of give all the right answers.

Who is going to pay for all the extra beds in local hospitals? Or special wards in local jails?

The goal of predictive policing is of course to prevent Columbine-style massacres and terrorist attacks, but it involves monitoring and profiling an entire population of individuals at all times.

Where do you set the filter? In other words, when do you cut it off as not serious enough and just ignore it?

And wouldn’t that cut-off itself be abused, in a particularly bigoted jurisdiction, to take all the wrong sort of people off the street, so nice people could ‘feel safe’ in their own neighbourhoods, or maybe to take one racial group off the streets so they would no longer compete for unskilled jobs with poor folks of the dominant race?

That’s already being done now, isn’t it, in some jurisdictions?

In my opinion the best cops have no hate in them, no bigotry, no prejudice. But there’s nothing to stop a bigot from joining the force and working his way up in it.

There's nothing to stop a bigot from running for sheriff or being elected governor, or even president.

No one has ever successfully managed to legislate for enlightenment, but then, no one has ever successfully legislated against prejudice and bigotry.

***

Preventive policing might even work, in that you would get arrests, and in the case of would-be terrorists, you might even find a truckload of explosive all ready to go, and a group or individual all set to carry out some plan.

In that sense, it would have been a success. That success would get high praise in the media.

Google has launched semantic search,** and I just read Facebook*** is doing heavy research into artificial intelligence, using the vast quantities of data they have gathered from us, quite frankly.

Semantics is the analysis of meaning. Artificial intelligence would use semantics to determine meaning, and with all the world now wired through a number of networks, using our phones, our devices, computers, automobile navigation systems, surveillance cameras, and a whole host of other sources of information, artificial intelligence will be used in preventive policing because it would require an almost infinite amount of manpower just to crunch the numbers and interpret data.

Preventive policing requires software and computer time, lots of it.

While a light might come on or a buzzer might sound on some police dispatcher’s board somewhere when Edwin tripped the threshold on his own personal algorithm, it is not quite clear whether the local police would consider it a high priority.

As long as people still had rights, they could always get a lawyer after some period of incarceration, or ‘observation,’ or even ‘treatment,’ and come back with a successful suit in a court of law.

My personal opinion is that the enabling legislation would have thought of that too—and done whatever was necessary to insulate the authorities from excessive responsibility for any mistakes that are made, or the inevitable civil and human rights violations that will surely occur.

But when you realize that most at-risk people really don’t have the resources to defend themselves in the first place, nor the resources to come back later, nor even to appeal ‘a wrongful conviction’ while they sit in a jail and rot—how in the hell that would ever be proven is also a good question—then a vast prison population composed of ‘at-risk’ individuals like Edwin doesn’t seem all that far-fetched.

It is almost a law of technology that all really revolutionary technologies bring disruption, they cause great and often unforeseen changes in the social context.

The infrastructure is already in place. It’s just a matter of time before this happens to some extent.

END

*Are police required to notify future victims?

What about potential future perps? Would a record of warnings or tickets be kept, and of course wouldn’t that also bear on the future outcome of a charge of ‘being at risk of committing a future breach of statute law?’

We got us a real can of worms here, ladies and gentlemen.

**Semantic search tries to predict the subject’s intentions, which of course has wider applications.

***Artificial intelligence would be used to draw conclusions based upon semantics, which may be defined as meaning, and multiple layers of deeper meaning. In the psychological sense, social theory would be used to define ‘at-risk’ indicative factors in any one person’s algorithm based upon past statistical analyses of individuals within social groups.

When these theories are based both on statistics and bigotry, ‘poverty breeds crime,’ for example, the possibility of abuse arises.