AI, Machine Learning and Analytics: Death of the journalist/artist
Once upon a time, machine learning consisted of research driven by simple algorithms. Over time, this has developed vastly from the very early Bayes theorem to now helping to minimize the work of journalists and even compete for Pulitzer prizes.
The Guardian reported last year that AI now has the ability to write ‘reading, flowing copy and churn out repetitive articles’. This is not AI’s first inroad into journalism either. Despite research findings suggesting that journalism will be one of the least affected professions by AI, it’s still impacting the content we are seeing online.
Clickbait can be generated by machines who pull data from over the internet, encompassing what readers respond well to and then focuses on producing content around this. We all seen the recurring “Ten things you didn’t know about Game of Thrones” type articles on a regular basis and machine learning is the driving factor behind this.
A Norwegian developer designed his very own ‘clickbait generator’, which scans the internet for clickbait articles and teaches his generator how to produce them. The system can then produce a single word before using something called Recurrent Neural Networks to help predict related words to form a headline. For example, after typing ‘Barack Obama Says’ into the generator, one of the completions was ‘Barack Obama Says GOP Needs To Be Key To New Immigration Policy’. Obviously, it isn’t grammatically correct or completely factual, but in theory, the headline isn’t a million miles away from being usable.
The New York times and BBC are reportedly making life much easier for their journalists by using machine learning tools to source and add additional metadata to audio and photos. This is just the beginning too, with plans to enhance headline writing and tags surely ready to make their mark too. This is a stark reminder that even the most creative jobs are being influenced by machine learning and the impact of this is set to continue.
Changing the face of art
In the past couple of years, there has been more and more digital paintings appearing online which have been drawn by AI. Google’s AI labs created a ‘Deep Dream’ system, turning images into a psychedelic rollercoaster, with the pictures being composed and decorated of primarily dogs.
In 2015, a team of researchers at the University of Tubingen in Germany created a system that would work in a similar vein. They taught their system to understand different artists stroke, colour and shape techniques so that it could essentially inspire the system to paint an image based on the selected artist. An image could then be run through the system, with a digital painting looking as ‘human-like’ as you’re likely to ever see coming out the other end of the conveyor belt.
Some of the results are staggering and more recently a digital image of Robin Williams shocked everyone when it was revealed that it had, in fact, been drawn by a computer. The system perfectly captured the emotions of the late actor, conveying a sort of sadness that stuck with those who had seen it.
It’s only a matter of time before technology will reach the standards of award-winning art. Some artists believe that one day AI will be behind Pulitzer Prize winning journalism. It seems almost inevitable that if given the resources, programming and enough human input, that AI will eventually be able to perfectly capture the emotions of art to construct something not even humans have the capability of creating.
