Home / Blog / The fading line between “fake” and “true”

The fading line between “fake” and “true”

I took this picture last week end in a Turin park. I like it, yet it is a fake on several levels …

Philosophers have been debating for thousands of years the concept of “truth”. You’ll probably remember Protagoras of Abdera criteria:

“Of all things the measure is Man, of the things that are, that they are, and of the things that are not, that they are not”

often captured by the sentence: “Man is the Measure of All Things“. A lot happened since Protagoras and although physicists would not buy into that statement (preferring that we can only discuss truth based on independent observations) for most people that statement still applies.

Unfortunately, our own tools are distorting our vision of reality and give us the means to distort it even further to the point that the separation between fake and true fades away.

This thought was prompted by the announcement by Adobe on their new Super Resolution features in Photoshop. Photoshop is probably the most reputable fake generator of them all, reputable because it masks the “fake” under the banner of better “truth”.

Before going into the Super Resolution let’s consider the photo I took during a walk in a Turin park last weekend. Bright yellow flowers in bloom against a backdrop of leaves-less trees and a blue sky.

I took that photo using the wide angle lenses in my smartphone. That and the fact I was well below the flowers created a distorted image with the trees in the background that seemed to bend. It gave an “unnatural” feeling to the picture so I used Photoshop to streamline the photo getting a much more natural image, quite similar to the one I remember.
Is this a fake? Or the fake was, actually, the image created by the wide angle lenses?

The latitude in the exposure of the smartphone camera is narrower than the one of my eyes/brain and this resulted in a paler sky and paler green of the grass. Again a few tweaks with Photoshop to darken the sky a bit increasing its “blueness” and a bit more saturation in the grass to have the green pop up. Is this creating a fake or just getting an image that is closer to my perception?

The photo on the left is the original one, the one on the right uses the sky taken at a different time and place.

One of the problem with the photo I took was the presence, on the upper left side, of a lamp-post. Just a few click with Photoshop and I removed that ugly lamp-post from the image, resulting in a nice, un-interrupted, blue sky. Is this a fake?

For sure this is s slippery slope. Little by little I am changing the photo into something else that is more pleasing to me but that is less and less connected to the reality.

As a further example consider the two photos on the side: the one on the left is the actual photo, the one on the right is an image using the first photo but with the sky replaced by one that better fits the whole, providing a more dynamic image. This second one is a fake, may be an acceptable fake.

Photoshop has plenty of tools that can alter an image in subtle or dramatic ways. It is a continuum an placing a dividing line separating what is fake from what is real is difficult to say the least. The problem starts, actually, at the very moment we capture the photo. What we get as an image depends on how we have set some parameters (exposure, white balance, saturation, ISO, …) on the camera and then it depends on the algorithms used by the camera chip to process the data provided by the sensor. More and more these algorithms are based on machine learning and AI. They have been trained by processing millions of photos to find the set of tweaking resulting in the most appealing image -not on the one having the highest fidelity!.

Super Resolution is just a further step in this blurred landscape of true vs fake.

Adobe has created an AI algorithm by analysing (through Machine Learning) millions of photos to learn what details look like. This knowledge is applied to increase 4 folds the resolution of the image. The algorithm looks at the image taken by the sensor and processed by the camera software and insert extra pixels basically “inventing” 3 pixels for every pixel created by the camera, thus generating, as an example, a 40 Mpixel image out of a 10 Mpixel image.

Let me stress once more that those extra 30 Mpixel did not exist in the original image, in a way you get a photo that for 3/4 is an artificial one, yet it is a credible one!  How would you call something that is derived from the real thing for 1/4 and the remaining 3/4 have been created out of the blue? A fake?

On the other hand one should consider that that 1/4 picked up by the camera sensor is a limitation of the sensor. That sensor is only able to pick up a limited range of the reality at a 10Mpixel resolution. Had the camera had a 40Mpixel sensor most likely the resulting image would have been pretty similar to the one generated by the Adobe Super Resolution software. So, in a way, one could say that the image resulting from the Adobe processing is closer to the reality, truer, than the one captured by the sensor: this is the fake one, not the “artificial” one!

See where am I going?

The tools we are using to transform the reality, the world of atoms, into digits, the Digital Transformation, are blurring the separation between real and fake.

On the other hand, as Plato pointed out, our brain (and senses) are also distorting the reality “out there” by creating an inner reality that in a way is “fake”. Nevertheless we think and act based on that “fake”. In this respect artificial intelligence is not that much different in our natural intelligence.

Interesting thoughts…

About Roberto Saracco

Roberto Saracco fell in love with technology and its implications long time ago. His background is in math and computer science. Until April 2017 he led the EIT Digital Italian Node and then was head of the Industrial Doctoral School of EIT Digital up to September 2018. Previously, up to December 2011 he was the Director of the Telecom Italia Future Centre in Venice, looking at the interplay of technology evolution, economics and society. At the turn of the century he led a World Bank-Infodev project to stimulate entrepreneurship in Latin America. He is a senior member of IEEE where he leads the Industry Advisory Board within the Future Directions Committee and co-chairs the Digital Reality Initiative. He teaches a Master course on Technology Forecasting and Market impact at the University of Trento. He has published over 100 papers in journals and magazines and 14 books.