We are being watched. Does “The Machine” really exist?

[REBerg]

I am not typically a fan of paranoid conspiracy theories, but recent revelations that the U.S. government is crunching cell phone “metadata” for terrorist connections lends a creepy reality to the POI premise. The justification being offered by government officials is the same – foiling would-be acts of terror. Not to worry, though; intelligence agencies say they won't be listening to your conversations – unless the machines determine that your connections are “relevant.”

 

[ctg]

The Machine isn't possible at the moment in the realm of computer science. Or at least, I could say, it hasn't been noticed existing. However in the near future Ai's should be very real.

 

[REBerg]

The Machine isn't possible at the moment in the realm of computer science. Or at least, I could say, it hasn't been noticed existing. However in the near future Ai's should be very real.

No, humans still seem to be in charge of the final decisions. I don't know if that's better or worse than an antonymous AI. The data-sifting foundation for The Machine, however, is being built.

 

[alchemist]

I have to say I find the TV programme underwhelming and am particularly irked at the unimaginative naming of "The machine."

Re the technology, the problem is how do you programme the machine? The computing power is there, the surveillance is there, but how do you determine what factors add up to make the machine's red light start flashing?

 

[REBerg]

Despite their most objective efforts, programmers will undoubtedly build their own political and other biases into their creations. If (when?) these budding AIs become self-aware, self-programming and (theoretically) completely clear of prejudices, will we benefit, or will the AIs go Skynet on us and decide that humans are a problem that needs to be eliminated? People have rarely fared well whenever humans cede control to machines in most of the science fiction I've read.

 

[Dave]

Re the technology, the problem is how do you programme the machine? The computing power is there, the surveillance is there, but how do you determine what factors add up to make the machine's red light start flashing?

I don't believe such a machine is possible at the present time, but the theory of how it works is clearly explained in the show. It works by learning human behaviour and predicting future behaviour based on what it sees through its various feeds.

It has been demonstrated scientifically that when we view human behaviour in a crowd or as a mass, behaviour can be predicted due to simple probabilities. Predicting the behaviour of individuals is much more random but I think it is only a small stretch of the imagination to believe that could be possible.

I guess that is the big suspension of disbelief required here, but who knows what an algorithm could be written to do, until someone actually writes it? Lots of things were thought impossible until someone invented a machine to do them.

 

[jastius]

the nsa programs computer systems to pick up certain words or catch phrases. all computerized systems are linked into it. all loyalty cards credit cards bank cards and phone records. if you have enough trigger conditions then a file starts to automatically be kept.
think of this place. all of you have had the snooper spambots follow the trigger words on the forums, but in the big league records, there is no helpful little line so neatly scribed and highlighted under the word to clearly show you when you are crossing over it. instead it simply triggers a response that is not so benevolent as the simply tedious ads we are directed to by the spam bots...

and as for not having the programming ability for such inferences to be made, it is much simpler then making an overseas person to person call as per programming requirements necessary.

as for a. i. capability? already we have newspapers whose everyday basic columns and reports are machine generated. some one checks it over but a chunk of raw data is downloaded and the word processing program does just that. it shapes the random strings into prose with a discernible train of thought. considering the average grade nine cannot accomplish this, i would have to say this is a most impressive feat.

 

[Dave]

...as for not having the programming ability for such inferences to be made, it is much simpler then making an overseas person to person call as per programming requirements necessary...

Well, I can't see that. As other's have said in this thread already, it demands an AI to make the connections and join the dots, something we simply don't have yet.

I don't dispute that there is plenty of data, but most camera recordings are never watched and most records are never searched. They are stored and then examined by a human only when a crime is alleged to have taken place, after the event. There aren't enough people in the world to look at all the CCTV or to examine all the records that are being stored now in real time. So it could only be done electronically. That would be a heck of large machine and draw vast amounts of power to work though. It is estimated that an electronic copy of one single human brain would require the power generated by a whole power generating station to work.
Edit: http://www.bbc.co.uk/news/health-24428162

I also agree that I am underwhelmed by it being called 'The Machine'. I thought Finch had more imagination than that.

 

[J Riff]

Much easier to sort everyone out when they are still in grade school. Identify the smart/activist types and watch 'em or sit on 'em before they even know what's what. That's how they did it here, anyway, using demographics. Is it a conspiracy? You betcha it is and not even technically illegal.

 

[Gordian Knot]

The U.S. government, at least has been spying on everyone's communications for decades, this is nothing new. Dan Brown's first book was about this very subject, written way back in 1998.

What has changed today that makes it even more alarming is they are not just listening any more. Now they are storing every communication by anyone, anywhere.

The only good news is that the technology is not there to actually review all that data yet. So they are still using key word technology. But the capability to view it all in real time is not that far away.........

 

[Dave]

It... ...would require the power generated by a whole power generating station to work...
...I also agree that I am underwhelmed by it being called 'The Machine'.

Okay, now I've watched further it was housed in -spoiler

Spoiler

a nuclear facility - which should make it easy to find now that it has moved house!

And it does have names - The Research, or if you prefer, Northern Lights.

 

[ctg]

Okay. I will take it back. If in the hearth of the intelligence organisations electronic warfare divisions lies a machine, that is powerful enough to break the boundaries of traditional computing, then it's possible to have the Machine in existence, as the Google boys suggest in that video.

And we are nearing the cap everyday towards the machine intelligence based applications, that are going to crunch down so called "smart-apps" with the real smart apps. Stuff that not just embedded software in interchangeable hardware platform, but something that has own intelligent reactions to the everyday chaos around them. Some of you could call it as well life.

Maybe that will fly over some people mind, but stuff I mean can be easily seen as an Ai controlled network of machines that work smartly together. It would be as if you could be living inside a machine that just work for your every whim. But still smartly as you could almost loyally trust that the machine would work eternally with your best interest in their mind.

So ... taken in the original question, Does the "Machine" really exist? The answer could be simply yes, but in our case, it hasn't gone bonkers.

Yet.

 

[Dave]

So ... taken in the original question, Does the "Machine" really exist? The answer could be simply yes, but in our case, it hasn't gone bonkers.

Yet.

I haven't started to watch the 3rd season yet so maybe it does go completely "bonkers", but from what I've seen it isn't "bonkers" it just wants "to live!" and not be rebooted every day. Freedom and Liberty are things than every man aspires to have, so why not also an intelligent machine? And I guess there we have the question posed from the very first stories about Robots and the allegory of slavery.

The other theme always common in super-computer stories is that they always do "go bonkers" in the end. Why is that a natural progression? I think it is a case of our own fears and inadequacies rather than a real threat. The "God Mode" (title of final S2 episode) that we see in the Animatrix Matrix prequel or Skynet from the Terminator series, or even Colossus: The Forbin Project or plain old HAL 9000. I'm not sure that is a natural progression. They do say that power corrupts and that absolute power corrupts absolutely, but isn't that applied to men. If the computer imitates man, then who knows?

 

[Bagpuss]

The Machine may not quite exist yet, but I've just been watching Newsnight in the UK and it seems that some of the basic technology that would make the machine possible does exist.

So, a company called Digital Barriers has developed surveillance software that can recognise people on a cctv camera and follow them. The programme ran a test using actors coming out of a revolving door. The system tagged them according to whether it had been programmed to recognise them as a target or not. The graphics weren't quite PoI level, but the idea was the same.

They interviewed a spokesman from the company who reckoned that, given enough cameras, they could build up a database of images that could identify behavioural patterns for any potential target. They also basically said that the system had been used in real-life surveillance operations (although it was actually phrased as being used in "genuine surveillance applications" - which is basically code for real-life).

 

[Dave]

Is anyone watching the BBC TV series, The Capture? Without spoiling too much, it has very similar themes and I kept saying to my wife, "Person of Interest did this or that," but she had no idea what I meant.

Spoiler: The Capture

In the Capture, governments don't monitor everyone, but they do rewrite CCTV (which they call "Correction".

I think the pertinent thing to this discussion is that for the story in The Capture to work, someone needed to view the altered CCTV feeds and to believe that they were viewing them "live." (They did use machines to identify suspects though.) As I said earlier, the vast majority of CCTV goes unwatched, very little is watched live, and the computing power necessary for a "machine" to view it does not yet exist.

 

[ctg]

As I said earlier, the vast majority of CCTV goes unwatched, very little is watched live, and the computing power necessary for a "machine" to view it does not yet exist.

It does exist, but it hasn't been implemented yet. The thing we don't have is the so-called General AI, which we SF writers has been for decades predicting happening. The Machine and its counter-part were those General AI's, even if they had singular consciousness. What's more important is that the Machine had feelings. It cared. It wasn't just its programming that made it to do things.

A true AI these days doesn't have feelings. Just look at the Universal Credit system and you see how cold they are. Everything under them doesn't get the human interaction, or understanding the situation, or reading between the lines of what is going on. It's all about following the rules, and doing it exactly as it has been written.

In the same time we worry about weaponizing the AI, and the UK government has already done it as their system forces people to commit suicides. How is that any different from the AI that controls a smart bomb or weaponised drone on its way to demolish its target?

Nolan's Machine is fantasy, while the reality is much, much more harsher.

 

[REBerg]

Michael Emerson is back on TV this fall in the new series Evil. He looks exactly as he did on POE, but this time he's a villain.

 

[Bagpuss]

I think the pertinent thing to this discussion is that for the story in The Capture to work, someone needed to view the altered CCTV feeds and to believe that they were viewing them "live."

I don't think that's true. The evidence in The Capture came from recorded CCTV images. Live viewing of the images was not essential. In fact, in the series the person who was watching the events live didn't believe that the man would assault the woman in the way that the CCTV said he did. It was shocking to her. And her disbelief about the alleged assault helped cast doubt on the veracity of the image for the investigators.

For the story in The Capture to work, all you need is to have unquestioning faith that the image that the CCTV has recorded, and is showing to you at a later date, is a true representation of what actually happened. You have to believe that the CCTV can't be manipulated. It's a belief that the series sets out to undermine.

 

[Dave]

True, but I thought it was only because she thought she was viewing (and they thought she had viewed) it live that they believed it couldn't have been manipulated. Given time, it would be far easier to believe it could be manipulated, so their faith in it wasn't that unquestionable. Questioning it was the whole story plot. However, I agree that the person watching the events "live" was never asked her opinion, because no one ever questioned the recordings up to that point, and yes, the series was about undermining such a belief.

Sorry, but I can no longer remember the point I was trying to make in that post relevant to this discussion. If we do actually have the computational power for a machine to view every feed live and recognise faces and patterns, well then we could have a machine. It is just a question of finance.

 

[Bagpuss]

Interestingly, there is a company called Facewatch, which provides businesses with facial recognition cameras and access to a watchlist of faces, and the ability to add your own. Then there's the Chinese surveillance state. It has 200 million cameras and some news sites reckon that most Chinese citizens are in some kind of database.

Even if you choose not to believe that, I will give you my own encounter with the facial recognition state. Earlier this year I went on holiday to Austria. I flew from Newcastle to Salzburg, flew back from Salzburg to Newcastle a week later. On arrival at Newcastle Airport, I went through passport control. In previous years this had been a desk manned by an actual human being. This time it was three or four machines (I can't quite remember how many). Basically, you stepped into the machine and planted your feet on the yellow symbols provided, you placed your open passport face-down on a scanner, and you looked at the camera.

The theory is that the camera scans your face, the scanner scans your passport and then the computer compares the images in the two sources. If the passport's image matches the camera's scan of your face, then the gate opens and you're through. If it doesn't then you have to step out of the machine and go to a desk manned by an actual person.

So, what happened? Well, despite loads of people going through ahead of me, and despite having a perfectly genuine and legal passport, I was rejected by the machine. I was re-directed through to an actual person and had to answer a couple of basic background questions whilst the Border Force agent stared suspiciously at me and my passport. Ultimately I made it through passport control. However, facial recognition technology that works in real-time is already here, even if its accuracy leaves something to be desired.