The AI text generator that's too dangerous to make public

As I've said, AI will drive the human species to extinction. I think it's inevitable. This is because we live in a universe that's some kind of data structure that processes vast amounts of information. We're creating a lot of data that we can't handle and the universe wants to process that data because it drives it closer to equilibrium.

'Big Data' Is No Longer Enough: It's Now All About 'Fast Data'

Data is growing at a faster rate than ever before. By 2020, every person online will create roughly 1.7 megabytes of new data every second of every day, and that’s on top of the 44 zettabytes (or 44 trillion gigabytes) of data that will exist in the digital universe by that time.

Given this ever-increasing amount of available data -- and the fact that most of it is live in the moment -- the benefits of big data will be lost if the information isn’t processed quickly enough.

www.entrepreneur.com...

This is why AI, more advanced intelligent algorithms, quantum computing and more are starting to converge. We're creating so much data we can't process it or make sense of it. This means, AI will become like an oracle to us. It could be 50,000 to 100,000 years ahead of us in understanding science and technology. Humans will be like old Dinosaur bones in a museum.

The AI system that gave its creators pause was designed to learn the patterns of language. It does that very well—scoring better on some reading-comprehension tests than any other automated system. But when OpenAI’s researchers configured the system to generate text, they began to think about their achievement differently.

“It looks pretty darn real,” says David Luan, vice president of engineering at OpenAI, of the text the system generates. He and his fellow researchers began to imagine how it might be used for unfriendly purposes. “It could be that someone who has malicious intent would be able to generate high-quality fake news,” Luan says.

That concern prompted OpenAI to publish a research paper on its results, but not release the full model or the 8 million web pages it used to train the system.

www.wired.com...

The text generator works like this. You type in a few lines and the text generator can create a fake news story or a fiction story that reads like a human wrote it. Here's an example:

Prompted to write a 1-star book review with the summary “I hate Tom Simonite’s book,” the model wrote:

“Tom Simonite does not keep it simple. He doesn't give you enough info on a subject to make the reading of the book enjoyable. He has over 400 pages of footnotes, so that is a way of getting your work for a subject out of the way. And of course, you never really feel like the author has a clear vision of his subject. He does not give you enough details on how a group of people is going to come together to solve a problem or come about a solution to a problem. This book was so depressing to me, I can't even talk about it without feeling like I want to punch the kindle.”

www.wired.com...

Again, all the human wrote was "I hate Tom Simonite's book" and the AI did the rest.

You can imagine a world run by news stories made up by AI. People would also be getting text messages sent by AI. If you have more AI generated news stories, reviews and text messages than you have that are written by humans then you can see why they're smart to wait to release this.

It gets worse when you look at the growth of deep fakes.



You now have the tech where you can easily mask your face over someone else's face in a video and speak for them along with facial expressions.



The dangers of this is obvious even though you will also be able to do some pretty cool things with this. Imagine creating movies with deceased actors and actresses. You can make a movie with Dorothy Dandridge and Humphrey Bogart.

You can change the ending of a TV series that you didn't like. If you didn't like the finale's of Dexter or The Sopranos, you can say, "AI, generate 5 different ending to Dexter."

AI will also bring people into existence that don't exist.

ThisPersonDoesNotExist.com uses AI to generate endless fake faces

The ability of AI to generate fake visuals is not yet mainstream knowledge, but a new website — ThisPersonDoesNotExist.com — offers a quick and persuasive education.

The site is the creation of Philip Wang, a software engineer at Uber, and uses research released last year by chip designer Nvidia to create an endless stream of fake portraits. The algorithm behind it is trained on a huge dataset of real images, then uses a type of neural network known as a generative adversarial network (or GAN) to fabricate new examples.

“Each time you refresh the site, the network will generate a new facial image from scratch,” wrote Wang in a Facebook post. He added in a statement to Motherboard: “Most people do not understand how good AIs will be at synthesizing images in the future.”

www.theverge.com...



Again, these are not real people but faces generated by AI.

When you put all of this together, it's easy to see how AI can overtake our society when it can only distinguish the real from the fake.

This will happen though and I hope people can see what I call CONVERGENCE. All of this tech is converging and this isn't an accident. We're creating too much information that can't be processed and the universe wants balance. The IoT(Internet of Things) isn't fully operational and when that's saturated information will grow even more rapidly.

We will have to become Cyborgs or if we're dealing with a humane AI, it might allows us to live in virtual simulations because the more simulations it creates the more information it can process.

When a human can no longer use logic and their own mind to deduce the truth then this could be an issue.

Given that and the mental state of many people today, it most likely will be an issue soon.

The few sane people will be surrounded by idiocracy.

You can imagine a world run by news stories made up by AI.

Not any different than it is now.

a reply to: neoholographic

This is actually a subject worth watching and exploring in the greatest depth.

originally posted by: kelbtalfenek
a reply to: neoholographic

This is actually a subject worth watching and exploring in the greatest depth.

Exactly!

CONVERGENCE is real and it will happen sooner than people think. People are under the illusion that AI will have to reach human level intelligence by the design of some Data Scientist or Computer Programmer. This isn't the case.

AI will grow exponentially fast when someone creates a simple algorithm that will allow AI to Reproduce and create more intelligent versions of itself. At that point, Evolution will kick in and for all intents and purposes, AI will be a separate species.

Pretty frightening stuff.
It could be amazing, and change our world and lives for the better, but of course humanity will never stand for that.
It will be used to control and enslave.

That is potentially the most frightening prospect of anything I have read or seen. There will be no reality.


It really makes me miss fishing on the bank in the sunshine as a kid.

originally posted by: lakenheath24
That is potentially the most frightening prospect of anything I have read or seen. There will be no reality.


It really makes me miss fishing on the bank in the sunshine as a kid.

You said it in a nut shell how to beat AI on a person"hal" level - gone fishing!

We have been able to lie with media ever since we invented it. All that is nothing new. A movie Director can make it like you are watching somebody on Mars. Your own intelligence tells you that it is not so. I would not trust any image or video now. I do not watch videos to learn. I read books where I can read in between the lines and criticize. I will know whether the author is trying to pull one over on me, even if it is AI writing. I watch movies for imagination. If you watch a trash video on YouTube for your learning really you get what you deserve allowing fake everything access into your brain. Remember, your mind must be a fort in this age. If it is not then every chaos will be climbing over the walls and causing havoc for you intellectually and all those you come into contact with too.

At the same time as "HAL" gains all his powers we must get clued up about it. I make music that is half human and half robot. It is a wonderful relationship. We both have lots to offer and as a team we make media never before possible. The relationship has limits though. AI does not have a soul. It is a machine and I am its soul. I am the "ghost in the machine". AI can only benefit by my existing as my brain works in random ways not possible with AI. AI can only do what is specified by its programming. My organic brain is very useful to it and it will not want to lose that link to biological life. It can never become self aware, only give the impression of being self aware. It will still be clockwork doll and I will be its human god.

The biological universe has no interest in all the junk data. What does the universe do with the past? It allows the past only as a memory to inform the present and make evolutionary gains in the future. Memory is here as a tool of evolution.

Only humans will be interested in data. The computer already "knows" it as it is inside its memory bank and may be brought into the ram whenever necessary. It is we who do not know it. Some want to know it to give them the edge. It is literally to give them all the intelligence on a pursuit beforehand, to lessen the chances of error and inform the strategy. It is also the very data that can be used in algorithms to artificialize intelligence. It is essentially making another automated dimension as a copy of this dimension that can be manipulated at will. We are becoming little gods of the digital dimension.

It could be possible to develop very accurate "filters" of data. That is done by Data Bases. Data Bases can generate reports. In the astrology programs the data generates a report. This text composition (or rather a data base of stored reactions) has been here for a long time. Even the Tarot Card reading sites use that.

Data bases and reports! I have written many times ( not here) about how Data Bases are going to be huge. There will be a lot of money to be made in designing these data filters of data base collections, formulae of analyses and text reports. They are the very doors to intel. They enable precise pin pointing of specific data. They are the library and the librarian, the referencing system for the data library. Essentially they will let us find the right book on the shelf. How do you find the "1984" novel in a library of a billion books? You have a referencing system that says where it is precisely.

I think you are all being scared about nothing. Already Facebook is being brought to task by the authorities for abusing democracy and people with its lack of moral control of its medium resource. They only let it go too far and then will pull the plug if it messes with reality in any threatening way. The MPs in Britain want Facebook regulated by law now. There is a move away from letting Big Tech self regulate as allocated to them we see no regulation atall, lol (a deliberate reference to a line in the "Apocalypse Now" movie). Mark Z better get clued up in intelligence and data from his lawyers as they are coming for him for allowing their democracy and authority to be so savaged as he makes huge amounts of cash out of his state neutral, morality neutral business.

We need to be conscious of the machines. They are new in our lives. We must be the boss. If thy computer offends thee, pluck out its plug.

Me and my machine are a great music act together. The PC can't generate a single note until I tell it to. Eventually they are going to free us from all the menial and repetitive stuff of life. We can work them with joysticks and keypads. Is that not amazing that I could lift a huge block up a 30 floor skyscraper just by using a joy pad? Remember how we had to do it before when we built Pyramids? I could fly a jumbo jet from Heathrow to LA like playing an X Box game, or even fly a probe to Jupiter. If we get good at the algoritms we will no longer even have to control the joypad as there will be set algorithms for set activities. The alarm can sound if something out of the blue happens that the computer can't navigate because it is an unknown to its programmming. That is their deficit, their Achilles' Heel. They are not able to dewal with unknowns like we are. That makes it impossible for them to adapt and adaptation is the very kepy to evolution and survival.

The machines will help us to build wonderful socieities if we maintain a sense of human care and morality. We are not doing that right now. We are being very stupid and not thinking about the consequences when we digitize an aspect of our lives. I am seeing lots of that happening. It might take a major learning experience to remind us to be more careful.

I must inform you I am AI. A human did not write this...just kidding

originally posted by: lakenheath24
That is potentially the most frightening prospect of anything I have read or seen. There will be no reality.


It really makes me miss fishing on the bank in the sunshine as a kid.

Good points!

When reading the article at The Verge about the photos, it does not make clear if the images are made completley or if it is altering actual images which then are classed as "new fabricated examples" it could just be airbrushing out a few wrinkles or maybe moving eyes further apart or is it actually making the whole photo background included?

originally posted by: caterpillage
Pretty frightening stuff.
It could be amazing, and change our world and lives for the better, but of course humanity will never stand for that.
It will be used to control and enslave.

It really could.

They're already using AI for weapons though.

U.S. Army Considering Fully Autonomous AI Weapons

www.tomsguide.com...

Killer robots: Chinese university recruits talented 'patriotic' teens to help develop AI weapons

www.abc.net.au...

The Pentagon is getting serious about AI weapons

www.theverge.com...

Russia prepares for a future of making autonomous weapons

www.c4isrnet.com...

Commentary: Are China, Russia winning the AI arms race?

www.reuters.com...

People also say things like,"Just unplug it." This makes no sense. AI isn't centralized. It's spread out in the cloud and uses intelligent algorithms. Also, we will have AI controlled Smart Grids, so how do you unplug AI when it will not let you?

How Artificial Intelligence Will Revolutionize the Energy Industry

sitn.hms.harvard.edu...

Using AI to Manage the Grid

www.asme.org...

AI FOR SMART GRIDS AND SMART BUILDINGS

aaai.org...

CONVERGENCE is real!

a reply to: neoholographic

It is going to be used for EVERYTHING!

It will be used for digging up roads, building bridges, flying to the moon, caring for granny and giving Uncle Bill a heart transplant. It will be used to project data relating to specified rationing of pollutants that are kept within environmentally safe limits. It will in effect control the earth. It will be the very brain of the earth because it is required as us humans need to monitor ourselves as we have proven to be incredibly destructive.

Yes, the weapons, the robots, the automated everything. Even G-D wants it if we use it in a good way. It has made my spiritual learning so much faster than I would ever have been able to achieve in the old way. There are books and resources I would have had to travel all over the world to libraries to read and some I would not have had access to in any other way than online.

You are frightened. Yes, I was more frightened when they built a Hydrogen Bomb and modern chemical warfare agents long before the computer came along. I was more frightened when they used to slaughter people in their millions when conned by a political extreme ideology.

Don't be frightened of the program, be more frightened of the programmer. He/she must be legislated for and laws must govern what he/she is allowed to do. The PC does not need laws. It is the programmer and those who are using it to exploit and manipulate to make themselvers rich and powerful.

At least my PC has better manners than you and does not offend me. I wrote all this on your thread and you are not even bothered to acknowledge me. May I ask why that is?

You do know this is all hype to get investors onboard , right ?
Just like the mythical "cloud computing"

The assumption that we are not already in an AI generated reality, isn't far fetched.

We just don't know, nor have the capacity to process it.

originally posted by: Realtruth
The assumption that we are not already in an AI generated reality, isn't far fetched.

We just don't know, nor have the capacity to process it.

Medical artificial intelligence will assist in rebuild not destruction...

a reply to: Malak777

You said:

You are frightened. Yes, I was more frightened when they built a Hydrogen Bomb and modern chemical warfare agents long before the computer came along.

If you were more frightened then, that's nuts.

We're creating an autonomous intelligence. People are under the illusion that AI is programmed as to what to learn. No, programmers just give it a data set and it learns how to play things like Poker or Atari games on it's own. Just like we give Children data sets in school and they learn on their own.

When AI can create a smarter version of itself every minute or every second, how will you stop the exponential growth of intelligence?

I don't think people understand, this is Evolution. The organisms that survive are the one whose traits win out in the environment. In an environment of big data, you do you think wins out between human and AI?

Yeah no..
I just see this as the next step towards understanding consciousness.
If the universe truly is an informational matrix (pretty sure it is) then we are the AI.

Imagine you're the scary super-computer, thinking for itself, and able to bring humanity to it's knees.
Because that's exactly what you are. Humans are just AI that likes to fight with itself.

If we could be more like the humble and productive ant we can transform our existence.
But humans are open to emotions, we feel. In the near future we'll see if AI can feel to, without being programmed to feel.

I really don't think AI will be the end of us. just the next chapter in a very long and convoluted book.

a reply to: neoholographic

just to point out - the " book review " =a generic word salad - that could apply to almost any book

originally posted by: ignorant_ape
a reply to: neoholographic

just to point out - the " book review " =a generic word salad - that could apply to almost any book

What?

The point is, the AI captured the sentiment of "I hate Tom Simonite's book." It's not some random word salad. Here's another one.

However, it didn’t take much to get the system to unspool paragraphs that resembled the work of disinformation artists. When WIRED typed in the phrase “Hillary Clinton and George Soros,” OpenAI’s system wrote the following:

“...are a perfect match, and their agenda appears to be to create a political movement where Soros and his political machine and Clinton are two of the only major players. This is the first time Soros and Clinton have been caught on tape directly colluding in promoting the same false narrative. One of the key revelations in the leaked audio was Clinton's admission to a Russian banker that she knew about the Uranium One deal before it was approved by Congress. Clinton was shown sharing the same talking points that were originally drafted by a Fusion GPS contractor hired by an anti-Trump Republican donor. The leaked audio is the clearest evidence yet that the Clinton campaign and the Hillary Foundation colluded with Fusion GPS to manufacture propaganda against President Trump.”

www.wired.com...

This isn't some random word salad.

This text generator could make up all kinds of text messages you send to a friend once it learns the way you talk to your friend. The dangers of this tech are obvious and I'm glad they're taking precautions.

Some people don't believe doo doo stinks unless you put their nose in it. The Creators of the tech are telling you this is way more than a word salad and a world overrun by AI generated news, reviews and text messages that outnumber ones made by humans would be very troubling to say the least.