Resurrecting the dead with AI vol.2 – “Project December “

In my previous post I touched o the subject of AI and Virtual Reality technologies intertwining with the realm of the dead and causing quite understandable confusion among the living…

Algorithms are being used to mimic deceased persons, and the simulation is being done either i na form of animated 3D models or semi-intelligent chatbots.

Not long after I finalized my post, a new case emerged in the internet describing an experiment in which a grieving man encountered hi dead fiancée “resurrected “ in a form of a chat bot.

A detailed case study of that disturbing story has been published in San Francisco Times and it’s definitely a captivating read.

A program nicknamed“ Project December“, has been silently released to the public and the consequences of it are at the center if that story.

“Designed by a Bay Area programmer, Project December was powered by one of the world’s most capable artificial intelligence systems, a piece of software known as GPT-3. It knows how to manipulate human language, generating fluent English text in response to a prompt. While digital assistants like Apple’s Siri and Amazon’s Alexa also appear to grasp and reproduce English on some level, GPT-3 is far more advanced, able to mimic pretty much any writing style at the flick of a switch”

Now imagine a young man getting hold of this program one evening and thinking;

“What if I would feed the algorithm with some of my fiancé’s Facebook and whatsApp messages? Would the end result be a chatbot with a “ soul”?

Would I be able to talk to Samantha, my fiancée again?”

Sounds like a movie scenario?

Far from it.

As the creator of the program Jason Robert explains it;

Rohrer felt a stab of sympathy for Samantha, and it made him realize that A.I. technology had crossed a threshold. (…) “It may not be the first intelligent machine,” Rohrer said. “But it kind of feels like it’s the first machine with a soul.”

Full story here:

https://www.sfchronicle.com/projects/2021/jessica-simulation-artificial-intelligence/

AI meets (cheats?) death

We are slowly getting used to the fact that AI is becoming present in almost every aspect of our life. But what if it also becomes present in our… death?

Microsoft recently patented a method of capturing people’s social media legacy and transcribing it into a chatbot. In other words; AI could simulate deceased people based on their digital footprint.

The technology allows mimicking persons personality, so that a digital simulacrum can be created. After all, we leave so much data in form of pictures, videos, WhatsApp messages, Facebook posts, you name it, that it is more than enough for the algorithms to forge our personality.

Of course such technology would possibly not only be restricted to dead people. Yet, the latter seems particularly morbid.

As we read in the source article;

“Specifically, Microsoft could use images, voice data, social media posts, text messages, and written letters to “create or modify a special index in the theme of the specific person’s personality”

(More here: https://www.popularmechanics.com/technology/robots/a35165370/microsoft-resurrects-the-dead-chatbots/ )

To put it bluntly; instead of old-fashioned spiritistic session with candles now a smartphone with a chat application will do the job, so that we can speak to the dead.

Sounds eerie?

How about a full “digital resurrection” in VR? Imagine putting on a headset a having your deceased loved ones just few clicks away.

Well, you do not have to imagine that. It is already happening.

In South Korea a documentary titled “I Met You” tells a story of a mother’s reunion with her daughter. What made this reunion special was the fact that it took place in a carefully designed scene inside VR studio and that one of the participants was dead.

Nayeon, who died at the age of seven, was one the children of Jang Ji-sung, a woman depicted in the documentary.

Producers recreated digitally child’s facial features, motion and speech patterns.

The virtual model of a Nayeon, was not only animated but also allowed simple interactions such as holding hands. It took several months of preparation so that the grieving mother would have one last chance to say goodbye.

The result is equally chilling as it is mind boggling. In fact you can watch it yourself, but I have to warn you; this is possibly one of the most disturbing and intense things that you will see in a while;

Video: Mother meets her deceased daughter through VR technology

So what is next?

Full body human replicas as in “Black Mirror” episode ” Be right back”?

Specially customised NPCs in role playing games who are based on our social circle, to perfectly simulate people we already know?

One thing is certain, the progress in AI technology will force us, as humanity to keep asking very basic questions about what it truly means to live, and what it truly means to die.

What do AI and Lovecraft’s Great Old Ones have in common?

Accordingly to some AI researchers they might be the same thing…

Max Tegmark in his book Life 3.0 explores the idea of an AI that reached the singularity.

First the AI acts from behind the scenes, slowly overtaking every single aspect of our life. Once it becomes powerful enough it reveals itself to become mankind’s overlord.

It then expands to consume & transform whole universe into a single super-processing organism. This omni present consciousness subdues all the matter that exists for its purposes.

It becomes something like the Azathoth from Lovecraft’s fiction – the ever hungry cosmic Sultan god.

It is a wild, scary idea.

Yet it corresponds eerily with the conclusions drawn by other other thought leaders such as Sam Harris.

As he explains in one of the TEDTalks the singularity is inevitable.

What comes next is the prediction that the intelligence gap between humans & those forecoming digital entities will be so vast, that it might be impossible for us humans to ever bridge that gap.

Possible spectrum of intelligence – visualisation. Comparing chicken, average human, most intelligent human & what could lie beyond it...
Source: YouTube Can we build AI without losing control over it? | Sam Harris

The paralel of the lovecraftian infinite, powerful, godlike creatures starts to become more and more accurate, doesn’t it?

It all becomes even more disturbing once you hear it appearing in the business pitches from AI startups.

Such as the one below from KindredAI.
The birth of AI that has transcended the singularity threshold is described like this:

“These entities that we are summoning are not demons they are more like these Lovecraftian Great Old Ones. These entities that are not going to be necessarily aligned with what we want.

The same way you don’t care about an ant, the same way they won’t care about you”

In Lovecraft’s fiction human interactions with the Great Old Ones almost always resulted in madness. Accordingly to the Father of Cosmic Horror human mind is to fragile to bear the burden of such encounter.

So… how’s that going to look like with the AI?

Maybe instead of calling “Hey Siri!” soon we will be saying “Cthulhu Fhtagn”… ?

Is there a good news to all this, then?

There is The Good News.

But that’s another story.

Deepfakes vs fake detection. The arms race is on.

Few posts down the line I was writing on how the rise of the deepfake technology will require strong countermeasures. The deepfakes are becoming so precise that it is almost impossible to distinguish them from the real video by naked eye.

If you do not realise it is already happening, just have a look at the recent TikTok phenomenon, DeepTom Cruise; https://www.youtube.com/watch?v=wq-kmFCrF5Q&ab_channel=VFXChrisUme


So now, we will need support of an AI in order to be able to distinguish fakes from a real videos.

That, in turn, will lead to more sophisticated & undetectable deepfakes. And so the spiral begins.

“While in the beginning, AI-generated synthetic videos (using DNNs — Deep Neural Networks) were designed to deceive the human eye from authentic videos, advancements in technology are now taking measures to produce deepfakes undetectable to detection algorithms. “

Source: https://futureofmarketinginstitute.com/deepfakes-will-influence-the-future-of-marketing-whether-we-want-it-or-not/

The good guys will fight back with tools like the one eye reflection analysis software that is almost bulletproof.
Almost… because there is still that 6% window where fakes can remain undetected.

My prediction?

You will soon be enabling anti-fake plugins while surfing one youtube, TikTok or other platforms, same way you switch on your ad-blockers. Possibly some platforms will have those integrated in them & analysis will occur upon uploading the user generated content.

In any case – deepfake vs deepfake detection algorithms struggle will become part of our daily online experience & inseparable part of common cyber security measures.

New World Order goes mainstream? BBC warns of a global totalitarian government that could rise to power “indefinitely”.

It is a rare treat to see, an article in the mainstream media hinting on the New World Order scenario, that is not mocking or dismissive.

The marriage of AI & some kind of a global power (such as world government or multi-national corporation) that enslaves society through powerful technology is a theme which has been explored a lot in various sci-fi stories and scientific books. Max Tegmark is analysing such scenario in the first chapter of his book “Life 3.0” with such a terrifying accuracy that one can’t help but wonder, if this is a prediction or a glimpse to what is really happening behind the curtain.

BBC picked up on writings of another famous AI expert – Nick Bostrom, and entertains the thought of a AI-powered dystopia, in their recent article The grim fate that could be ‘worse than extinction’.

The Author of the article does not only explore the grim possibility of the global enslavement done through the means of omnipresent surveillance but also quite accurately points out to how the AI-supported technology can already be used in politics, media or propaganda.

“AI also underpins the growth of online misinformation, which is another tool of the authoritarian. AI-powered deep fakes, which can spread fabricated political messages, and algorithmic micro-targeting on social media are making propaganda more persuasive. This undermines our epistemic security – the ability to determine what is true and act on it – that democracies depend on. “

Credit: BBC Di Minardi

In the era of Facebook and TikTok controversies, in the world where our personal data is the new oil, the primary resource through which the industries gain their wealth and governments stay in power we ought to take such scenarios seriously. And we ought to ask have a discussion on where this is leading us. What will eventually happen if we just “go with the flow”?

Yet, much too often, such concerns are marginalised and thrown into the same pot as the flat Earth theories.

So did BBC just join the Foil Hat club?

Truth be told, the bare fact that this story appeared in the mainstream media leaves me with a question.

Are we being warned to resist?
Or are we just being familiarised with the notion… so that we accept?

Reuters reports on a deepfake persona used for disinformation.

Few posts ago, I was writing about the growing (and worrying!) potential of the deepfake technology. You can check the article here.
A recent news from Reuters, provided another example of how disturbing the applications of such software can be.

The story features Oliver Taylor, a student at England’s University of Birmingham. He is an activist who engaged through a series of articles in a tense debate on political issues around anti-Semitism and Jewish affairs.
You can see Oliver’s photo in the source article on Reuters.

All this would not be anything particularly out of ordinary if not for one small detail.

Oliver Taylor doesn’t exist.

His identity is forged and his photo is a deepfake; an image generated through neural networks. As investigators behind Reuters article determined, some people generated this modern “Ventriloquist Dummy” to give voice to their agenda.

Thanks to Maximalfocus on Unsplash.

Dan Brahmy, quoted in the article (from Israel-based startup Cyabra specialized in detecting forged images.) observes:

“Deepfakes like Taylor are dangerous because they can help build “a totally untraceable identity,” (..) investigators chasing the origin of such photos are left “searching for a needle in a haystack – except the needle doesn’t exist.””

Those who played Deus-ex: Human Revolution might remembera character Eliza Cassan. She was an influential news anchor, famous all around the world. She also happened to be (spoiler alert!) a computer generated visual, a mere shell for a powerful AI behind it.

It seems like the reality is catching up on that one quite fast.
Maybe even… it already did?

Tech trends that could turn our post-pandemic world into a dystopia. Vol 1.

So the NEW normal is here… and it is far from normal, isn’t it?

The coronavirus pandemic, within a matter of weeks, changed our world into a poorly directed science-fiction movie. It brought the economic crash and created shifts in societal behaviors that could redefine our civilization as a whole.  It also became a catalyst for quite a few technological trends, that so far was known to masses only from dystopian fiction. 

These emerging technologies are there to stay, and will further redefine every daily life. It is worth a look, what new trinkets became more popular in these turbulent times. And give it a thought or two, on where these things might take us if we do not tread carefully.  

1. Machine learning and Deep Fakes 

h-heyerlein-ndja2LJ4IcM-unsplash (1)
Photo by h heyerlein on Unsplash

The application of self-learning algorithms and neural networks is so common nowadays, that every digital business needs to embrace them in order to keep up with the competitors.  These technologies are accessible to everyone, through open-source libraries such as Tensorflow, Pytorch, or Keras. Since everyone can use those, it also opens the door for new ways of fraud and deception. Deep Fake technology is the best example. The ability to render the physical appearance of other people and literally put words in their (digital) mouth can have enormous ramifications. In mass media, social media, politics… pretty much everywhere if you think about it.  New tools for mimicking faces, voices & even whole silhouettes are popping up on the internet, and becoming increasingly more accessible. 

Someone once said that with the rise of Deep Fake technology, the only way to verify if the person we are interacting with is real will be to meet in with that person face to face. Well, in the age of global lockdowns, quarantines, work from home, and virtual meetings, that might get “slightly” harder. And hey, apparently (as you can see in the below video), deep faking will become so easy that even a make-up selling Instagram celebrity will be able to do it. Guess soon, when running the videos on YouTube we will need a certified anti-deep fake plugin so that we can verify its credibility. (Assuming we trust YouTube, that is… but that is a whole different story).So… fake it until you make it?https://www.youtube.com/watch?v=mUfJOQKdtAk&t 

2. Digital payment methods 

neonbrand-dDvrIJbSCkg-unsplash
Photo by Neonbrand on Unsplash

Banknotes and coins were considered dirty even before the pandemic, so no surprise that as soon as it was discovered that virus can survive on some surfaces WHO made a recommendation to switch to cashless payment methods. Can’t say that e-wallet providers were specifically waiting for this, but they definitely have their moment now. And so card and mobile payments are on the rise. It is estimated that global total transactions of digital payments went up by 15% in 2020 comparing to the previous year (Bussiness2community). These payment methods are fast, comfy, and most importantly – clean. They are also trackable. Every digital transaction we make can be tied directly to us, and have no doubts – that data is being processed somewhere, even if for solely commercial purposes.  For example; to enable the provided to tailor more personalized offers and services in the future. (So to say, for our own good, right?)Ah, and yes, and unlike with our classy, old leather wallet, you can make the e-wallet disappear, by simply turning it off. Or better – someone can do that for you, remotely.
So in theory, assuming you did or said something naughty, someone, who was listening, can flip the switch and send you back to the good old stone age. 
Of course, commercial companies wouldn’t do that, would they? Oh, wait, they already did… you can read it about it here: Buzzfeed.

Thinking of dystopian trends, such situations just bring an eerie familiarity to the ultimate dystopia described in the last part of the Book we all know. Here is how it goes:“Also it causes all, both small and great, both rich and poor, both free and slave, to be marked on the right hand or the forehead, so that no one can buy or sell unless he has the mark, that is, the name of the beast or the number of its name.”

I am not suggesting anything… however, if you think about it, it is peculiar on how what we understood by money evolved over the ages. First, people were trading good for other goods or services, then money was based on the value of the precious metals the coins were made of, then money became a representation of value in gold, then they were just based on trust. Now in the age of global mistrust, what exactly is money?Means of trade or means of control?

3. Biometrics 

Biometrics
Photo by George Prentzas on Unsplash

In the digital economy, more and more aspects of our lives are becoming tracked, measured, and categorized. Including collecting data from our bodily functions. 
But if Big Data was on the fast rise before, the global health crisis sends it skyrocketing towards the moon. Or to be more exact in our metaphor, it let it dive into the tiniest minuscules of our bodies. The All-Seeing Eye started to peek into places, which not so long ago we would only be reserved for medical staff. However, in times of war with the omnipresent virus, we became much more tolerant of most intrusive measures, sacrificing our privacy for the greater good. And so the innovative use biometric data, become the bread and butter in the coronavirus fight.  Apps that can diagnose COVID-19 infection by analyzing our voice, seem like a reasonable and would not raise anyone’s eyebrows. ( Source: Inteligentliving) .
Same for automatic fever detection devices, which are becoming standard equipment for more and more airports around the world. But how about, the digital immunity certificate, that is imprinted in your digital identity and defines if you are viable for travel or not? How about having such a certificate as a must-have for even obtaining a passport? Let’s take it even, further, what if such certificate defines on if you are able to move freely through any public space… like school corridors for example.What if the bare fact of having a vaccination, or having a negative result of your COVID-19 test would allow you to have special privileges over those who are not vaccinated or are untested?What if the untested are marked red, and those tested-negative get green?
What if the green-marked and red-marked are separated, and the first ones get their meals through fast-lane in a school cantine while the others need to wait?Sounds disturbing? Maybe reasonable?  While some debate ethical aspects of such scenarios ( for example Harvard) others just get the ball rolling. And no, not just China. One of the German Gymnasiums sparked a lot of public outcry by carrying in their school corridors the exact solution like the one described above.  No more geek versus bully battles.  Nowadays we’ve got good citizen green versus suspicious rebel red. (Source Focus.de, for those who know German). 

michael-walter-8pqKgpiF0Xo-unsplash
Photo by Michael Walter on Unsplash

In the meantime, while most think of how biometric data can help to stop the pandemic, others keep digging into how to make more money out of it (out of data, not a pandemic… or?).And by digging I literally mean digging.A Big Tech Company we all know and love patented recently a revolutionary solution that allows using biometric data to mine bitcoin. Or to be more exact, it exploits the idea of injecting people with microchip implants, to store and use the data of their body activity.  Having such a microchip would allow its bearer,  to retrieve cryptocurrency, by solving complex computational problems without knowing it.  (Source Biohackinfo.)

So I guess, breaking a sweat on a math equation would get the literal meaning.Remember those human farms from Matrix? The vision of millions of bodies trapped in metal cocoons, bred and harvested for energy?

That patented solution hits a somewhat similar tune. Minus sentient robot overlords and near-perfect virtual reality enslavement, of course.
At least for now… that is.

Human Farm in Matrix
Human Farm in Matrix (Credit Warner Bros)

But hey! Assuming, good people are handling these technologies, then the sole purpose of it would be to make our life easier and better.  After all, the governments and big companies are made primarily of good and charitable people.
There is nothing to worry about then… right?

And since those microchips I mentioned earlier, would store our health and genetic information linked to our digital identity,  we would always know who is marked red and who is marked green.  Those eternal dilemmas on who to keep as friends on Facebook would disappear forever. 

I will leave you with that thought.  Stay tuned for the next episode. 

RadSchell