#the creativity code

My book review:

In The Creativity Code, Marcus du Sautoy (a University of Oxford mathematician) dives into an exploration between the lines of artificial intelligence and human creativity through stories and examples on games, art, music, storytelling, and songwriting to name a few; the history of machine learning algorithms; and if programmed machines can develop a form of consciousness, producing this human trait of creativity, or just give an illusion of it through imitation. 

Here’s some excerpts:

"Fake” Artists on Spotify

[François] Pachet has been poached from Sony Labs and is now working for Spotify. Given that rumors have been circulating that Spotify is creating playlists full of songs by "fake" artists, the move is an interesting one. Music critics spotted a number of artists on Spotify who were notching up an extraordinary number of hits thanks to their inclusion on popular playlists curated by Spotify for meditation or running. A band called Deep Watch had recorded 4.5 million plays over a five-month period. 

When critics tried to find out who these artists were, they kept drawing blanks: no presence anywhere else on the internet; no upcoming concerts; no details anywhere of such a band. A rumour began to circulate that the music was being generated by ‘fake artists’ so that Spotify wouldn’t have to pay royalties. Spotify hit back: ‘We do not and never have created “fake” artists and put them on Spotify playlists. Categorically untrue, full stop.’ But it does appear that they were specifically commissioning minor artists to create songs under fake names at royalty rates that were much more favourable to the company than its standard deals with record labels.

Harry Potter and the Deathly Botnik

Machine learning is changing that. It’s now possible for an algorithm to take an author’s entire opus and learn something about the way they write. If they favour a particular word, there may be a high probability that this word will be followed by certain other words. By building up a probabilistic picture of how an author uses words, an algorithm could start to generate the continuation of a text. This is how predictive texting works. The literary results have been both revealing and entertaining.

This use of machine learning to create new literature has been championed by a group that calls itself Botnik. [...] The group has taken Seinfeld scripts and produced new episodes based on a mathematical analysis of past dialogue and even got an actor from Scrubs, Zach Braff, to perform a monologue.

Probably their most successful output to date came from training Botnik on the seven volumes of Harry Potter. The three pages it generated have a very convincing ring to them.

Magic: it was something that Harry Potter thought was very good. Leathery sheets of rain lashed at Harry’s ghost as he walked across the grounds towards the castle.

But there are moments of pure genius that could only have come from an algorithm:

Ron was standing there and doing a kind of frenzied tap dance. He saw Harry and immediately began to eat Hermione’s family. Ron’s Ron shirt was just as bad as Ron himself.

I guess for fans who are really desperate for more from the wizarding world this may be better than nothing, but it’s pretty plot-free and is unlikely to sustain much drama beyond three pages.”

(End of the excerpts)

Can you guess which painting out of the two below was made by an AI to mimic the art of Rembrandt?

Did you pick?

Are you sure with your choice?

The answer is the first one. (Did you spot Rembrandt’s signature on the second one?)