The Cheat Sheet

AI on song and look who’s facing the music!

Jinoy Jose P | Updated on November 28, 2019 Published on November 28, 2019

Attempting hip-hop lyrics ?

How I wish! But I’m not cut out for the job. And the headline was my feeble attempt to capture your attention so I can talk to you about Yona.

Who’s Yona?

Yona is a young writer. A report in Bloomberg told me she loves reading Margaret Atwood and articles about teen life. She sings about loneliness and relationships on her newly released track.

Sounds like a K-Pop star; I hope she won’t hurt herself as many of them do these days.

Oh dear, no. Yona is a robot.

Robot, and a writer?

You heard me. Yona writes songs. She has been created by London-based company Auxuman, which has trained the artificially intelligent machine to enjoy music, literature and write songs and more. Yona, the report says, can even read comments from her fans on her music posted online and learn from the feedback.

Well, I don’t know if I should be happy or be worried...

Well, that’s a debate we can reserve for a rainy day. Now, the arrival of robots and artificially intelligent programs such as Yona or Sony’s FlowMachines (which learns music styles by scanning songs) poses interesting questions around the way AI disrupts creativity and how the process can impact industries that deal with creative content, such as music, literature and painting.

That reminds me of a painter robot I read about some time ago.

Yes, Ai-Da, arguably the world’s first robot artist to stage an exhibition. The humanoid can do a portrait of a person just by looking at her face, can paint abstract works loaded with ideology and can even sculpt a bit.

Her creators claim that her works are as good as most human painters of today.

Impressive!

Clearly, programs such as Yona or Ai-Da (named after Ada Lovelace, the legendary mathematician) are pointers towards a future where AI could pose interesting challenges to creativity and art. For instance, how would you be enjoying or reacting to a novel that is written by a robot? How will one review a work of art that is created by an artificially intelligent program? Even more interesting are questions around the way companies and creators are going to price such products.

This is getting crazy!

Indeed. Take Yona from Auxuman (stands for auxiliary human). The company, which has also created similar programs such as Mony, Zoya, Hexe and Gemini, is aiming to extend the frontiers of AI music far and wide by introducing serious works that can match human talent. Auxuman released its debut album on September 27. According to Ash Koosha, an Iranian electronic composer and co-founder of Auxuman, the company is planning to present virtual entities that “give to you, what many humans, in many cases, go through a lot of mental turmoil to create.”

Wow!

Well, that statement is tricky. In fact, in the world of creative works, one of the most defining factors that determined the quality of artwork was the physical and mental troubles the artist would have undergone before s/he produced the said work. For instance, many suggest that Leonardo da Vinci took anywhere between four and 12 years to paint Mona Lisa. The painter went through several personal hardships before he finally finished the work of art. And that is one of the key factors behind the global acclaim the master artwork has commanded over the years.

Interesting!

Now, the moot philosophical question about an AI painter is (though the query may sound silly at this point in history when AI is just making its baby steps) that how will you appreciate and even price an artwork that can match the artistic brilliance and production quality of a Mona Lisa, or any similar great piece of art, if it has been produced by an AI program or a humanoid powered by machine intelligence.

That’s a big ask!

Indeed. Also, there lurks a danger. If you remember, when the likes of Elon Musk (of Tesla) openly toyed with the fancy idea of robot authors a few years ago, as part of the non-profit lab OpenAI (supported by Musk etc), many authors had warned that the language model OpenAI was developing (called GPT-2) was too dangerous for them “to release into the wild” because they feared it could create “deepfakes for text”. Evidently, a world where AI could write like George Orwell or Ernest Hemingway or a hybrid of 100 such authors, or create satisfyingly original fakes of such creative writing, the stakes are high not only for publishers or content creators but for connoisseurs of art as well because such phenomena questions the very idea of originality.

A weekly column that helps you ask the right questions

Published on November 28, 2019
  1. Comments will be moderated by The Hindu Business Line editorial team.
  2. Comments that are abusive, personal, incendiary or irrelevant cannot be published.
  3. Please write complete sentences. Do not type comments in all capital letters, or in all lower case letters, or using abbreviated text. (example: u cannot substitute for you, d is not 'the', n is not 'and').
  4. We may remove hyperlinks within comments.
  5. Please use a genuine email ID and provide your name, to avoid rejection.