Searching for Sanity

The disruptive technology of Artificial Intelligence (AI) is being increasingly tested and deployed throughout different technological sectors.
One area of particular interest to me is the use of AI in search engines. Google forums are awash with speculation on how the new search algorithm AI, amusingly named “Parsey McParseface” is used to deliver fast, accurate results. The problem according to many in the webmaster community; those results are don’t make much less sense. For many veteran webmasters, this seems to be a break with their longtime logical reality. Once you see these examples, you’ll better understand why some of the results might be so strange.  Hang on to your hats!

Keep in mind that many of these examples were generated with the gain turned up to 11. This begs the question of where set the AI network so it produces the best “sane” results.

Psychedelic or Psychotic?

As a demonstration of what AI actually “sees” in it’s massive parallel processing brain, the image below started out as a cloudy blue sky processed by an AI network trained to look for buildings, specifically pagodas. The results are surrealistic enough to make Salvador Dali jealous.

.

via GIPHY

How Artificial Intelligence Views The World

In these examples, AI networks trained to look for specific objects are processing ordinary images. The results are both humorous and shockingly strange views of our world.

Bizarro Processing

three examples of ai neural netwrks finding exactly what they are trained to find in ordinary images even if those items are not actually contained in the image

Artificial Intelligence Neural Networks find exactly what the are trained to find, no matter the subject image.

Here is an interesting view of artificial neural networks interpretation of ordinary images. It seems once trained, they  focus in single minded fashion on only what they want to find. This could lead to many cases of false positive identification in the application on SERP (Search Engine Results Pages). I’m sure there is much more to it.

 

More odd processing…in this case, the AI is looking at a photo of a blue sky searching for animals, progressively building complicated layers and eventually deciding on these strange new hybrid creatures.

four examples of AI networks idetifying strange animals from a photo of a clear blue sky

Strange “animals” found in a photo of a clear blue sky.

While these images where generated out of static noise.

16 odd photos generated from static noise by artificial intelligence neural networks

Neural networks identify images from static noise source images

Generating It’s Own Version of Reality

In the following video, discover how fast evolving AI is learning to generate images based on mere descriptions. This is where AI could very well be crossing into web search by creating a representation of the site text and making a value determination based on that image.

What Does It All Mean?

In the Coldfusion TV video above it is evident that AI is taking even Google co-founder Sergey Brin by surprise. If the advantage of using artificial intelligence is restricted to only a few and is used against society for profit, then the result will be grim. However, if it is put into the hands of everyone, perhaps the world could finally work together for the benefit of all. The question is : “will those who create the technology allow that to happen, or use it solely for their own ends?”

Disruptive so far?

Outside the lab, AI will be a very disruptive technology. In the fictitious Hollywood version, replacing humans with machines has had a long history of tragic results. 2001 A Space Odyssey, the Terminator series, you get the picture. There are clearly examples beginning to manifest themselves in the real world. Google using AI to maximize profits would destroy countless incomes for real humans. Do the AI creators care?

In the BBC series “Secrets of Silicon Valley” series 1, episode 1 of 2 – The Disruptors, Jamie Bartlett uncovers disturbing trends in the current application of Artificial Intelligence used by some online businesses. Here is the synopses:

” Jamie Bartlett uncovers the dark reality behind Silicon Valley’s glittering promise to build a better world. The tech gods believe progress is powered by technology tearing up the world as it is – a process they call disruption. He visits Uber’s lavish offices in San Francisco and hears how the company believes it is improving our cities. But in Hyderabad in India, Jamie sees for himself the human consequences of Uber’s utopian vision – drivers driven to suicide over falling earnings. Riding shotgun in a truck as it drives itself for more than a hundred miles on a highway, Jamie asks what the next wave of Silicon Valley’s global disruption – the automation of millions of jobs – will mean for all of us. In search of answers, he gets a warning from an artificial intelligence pioneer who is replacing doctors with software – an economic shock is coming, faster than any of us have realised. Jamie’s journey ends in the remote island hideout of a former Facebook executive who has armed himself with a gun because he fears this new industrial revolution could lead to social breakdown and the collapse of capitalism.”

Moving forward:

It all depends upon how artificial intelligence is used, for or against mankind. The photo examples above simply demonstrate how the technology used improperly could produce strange side effects in-compatible with humanity. However, tuned back to a sane level, AI could positively benefit all mankind.

If you are interested in helping ensure that Artificial Intelligence falls into the right hands, visit the Future Life Institute

Please feel free to comment below.

Further reading:

Stephen Hawing : Artificial Intelligence Could End Human Race

Microsoft pulls the plug on Tay AI after Twitter users train it to make racist remarks.

image credits – Google, Business Insider