Generative AI does not kill creativity!

Last month I read an article published in an Indian national newspaper written in Bengali on the impact of AI on the creative industry. The author of the article is a very well-known personality from India who has won several awards for a number of his musical compositions over the years. He is also a very accomplished singer and writer and he happens to be my childhood friend Anupam Roy. You can find a long list of all of his achievements on his Wiki page Anupam Roy. When I read the article, which he wrote in our mother tongue of Bengali, and received lots of reactions from his readers and numerous fans, the first thought I had was that I needed to discuss it with Anupam, and aptly translate and rewrite it in English. You can read the original version of the article here: মা নুষনু মা নুষনুকে আগে ও ঠকি য়ে ছে , ভবি ষ্যতে ও ঠকা বে , এআই কী দো ষ কর ল?

Anupam wrote about the impact of AI on his field of work. This was his reaction to someone using AI to sing one of his most popular Bengali songs “Boundule Ghuri” in the voice of one of the most famous Bengali singers ever, Hemanta Mukhopadhyay - a legend in his own right. Anupam and I have grown up together and agreed/disagreed over many things in life. Here we are working in two very different fields, yet our thoughts superbly align on the topic of AI. As a person who works in AI and as a co-founder of an AI startup I face many questions on whether AI poses a threat to creativity or not from lots of people. So what better way to answer some of these questions than to get some help from one of the most creative people I know!

The next challenge was how to best represent Anupam’s powerful Bengali diction and prose and effectively translate it into English. I wanted to be able to best reflect the emotion in the article, the relevance of the examples he has used to make his points, and depict a creative artist's view on using AI to recreate one of his works of art. Surprise surprise, I had to use AI to help me out with that (I will explain in a different blog how that process worked). The full translated version of Anupam’s article is at the end of this blog, but here are my observations on some of the points that caught my attention the most.

Let’s start with the title of the article:

Humans have deceived each other in the past, and will continue to do so in the future: Is AI really to blame?

By starting with this question Anupam alludes to the fact that consumption of “fake” content is not uncommon (forged money, art, fake reality TV manifested through Instagram reels, etc etc) but that doesn’t mean it’s not creative. Yes whether it is right or wrong is a different question altogether with a very clear answer of course! But in the mission to blame a disruptive technology like AI, people are re-packaging something that has been ongoing and has not changed.

As he positions his argument, he asks a very common yet important question which is often very carelessly answered by many but not by him. He says:

Now, regarding AI: is it good or bad? Such a question misses the mark! The perennial battle is between conservatives and liberals, the traditionalists and the innovators. I side with innovation. AI has arrived. It will change a lot overnight, and neither you nor I can stop it.

I say, neither you nor I can stop it but both you and I can build the future with it. This of course stretches way beyond the creatives and creativity, this applies to all of us!

Anupam then talks about the use of AI which sparked this commentary:

But back to the subject: my song "Boundule Ghuri," brought to life with Hemanta's AI-generated voice, is stirring discussions. Is this an experiment in sophisticated artistic engagement? Certainly not. Yet, it's an attempt at something new.

I think he wonderfully touches on the power of creating something new here with the help of a human-built technology. He talks about serendipity! Serendipity binds together knowledge, inquisition, and creative thinking to drive innovation. AI has shattered the barrier to entry by making this happen through the easy-to-use multi-modal solutions that have been made available to all of us. Harnessing this power of AI can and will help us unlock things we couldn’t possibly have imagined before. Check out the following article if you are interested in finding out more about why organisational innovation is highly dependent on increasing serendipitous occurrences at work.

Innovation at work: the power of serendipity - News & insight - Cambridge Judge Business School

He moves on to provide a powerful argument around the question - Are we seeing a shift or repurposing of creativity from creative output to creative input? He says:

Designing with AI thus requires a prompt; without quality input, AI cannot turn dross into gold. Hence, a talented Hemanta-like singer is needed before AI can refine and convey nuances. This is where talent and effort diverge.

An interesting point very well made. We need talent, novelty, and true creative thinking to create new ideas and new content on which AI can be trained and produce even better outputs in the future. Human creativity and new types of content will inevitably need to provide the fuel for continuous improvement.

Finally, I love the way he finishes his article:

Let AI present us with challenges. I welcome them.

If creativity is the tendency to generate or recognise ideas, alternatives, or possibilities that may be useful, then we should be more concerned with what is useful, than about the sanctity of creativity. True human creativity will cope perfectly with whatever is thrown at it.

I have learned a lot from him over the years as he was always up for any new challenge. I think we have a lot to learn from what he is saying to us here. We need to understand how we need to look at making the most of this unique challenge that AI has given us to use it to shape and build the future.

AI is here to support creatives like Anupam, and if others accept it with open arms like he has, we have a very exciting future ahead of us!

(The full translated version of the article is provided below)

Humans have deceived each other in the past, and will continue to do so in the future: Is AI really to blame?

Whenever something new emerges, a faction predictably begins their outcry. We've witnessed considerable protests (by the Indian Labour Union and others) during the 1980s when computers were introduced in India, with numerous trade unions in West Bengal staunchly opposing their use (in banking and other sectors), though these initiatives were government-supported. Buddhadeb Bhattacharya later acknowledged that such resistance was a historic mistake. Bengal could have become a Silicon Valley, but it didn't happen. Our youth could have found employment here, but they didn't. There was no need for migration to Bengaluru or Hyderabad. Yet, we failed. And now, years later, Bengal still struggles to stand tall in development, its face marred by incidents as shameful as those in Sandeshkhali.

But let me not wander off topic. Now, regarding AI: is it good or bad? Such a question misses the mark! The perennial battle is between conservatives and liberals, the traditionalists and the innovators. I side with innovation. AI has arrived. It will change a lot overnight, and neither you nor I can stop it. Attempting to do so would be folly. Our nation's children are making significant strides with AI globally, working with giants like Microsoft, Google, Meta, and others. Are Bengalis aware of how much NVIDIA's stock rose last year? Caught up in trivial debates on Facebook, many overlook these developments, remaining oblivious to the world's progress, still preoccupied with converting farmland into fishponds, while openly rigging elections in Chandigarh and getting caught on CCTV. Apologies, I digress again. We shy away from discerning truth from falsehood, allowing our intellects to petrify. But why all this discussion? My song "Boundule Ghuri," featuring a voice akin to Hemanta Mukherjee's, is gaining traction on social media. Naturally, opinions are divided. I shared it, too, noting it as "interesting"—a deliberately chosen adjective. Whether the song is good or bad is not the crux of the matter. No one claims Hemanta himself sang it; that's beside the point. It's explicitly mentioned that AI was used to replicate Hemanta's voice. Personally, I find the fuss over AI in music somewhat trivial. My concern deepens when considering criminals and propagandists exploiting AI. Imagine a doctored video of Netaji advocating for a temple in Delhi going viral on WhatsApp, misleading thousands.

We avoid verifying truths and falsehoods, turning our brains to stone. But back to the subject: my song "Boundule Ghuri," brought to life with Hemanta's AI-generated voice, is stirring discussions. Is this an experiment in sophisticated artistic engagement? Certainly not. Yet, it's an attempt at something new. The question of artistic quality aside, did Hemanta wish for this? Unknown. It was done without permission. A crime? Unclear. That's not the point of our discussion. Someone tried, and indeed, it's a commendable effort. AI needs a human source for accurate imitation. The better the mimicry, the more precise AI's output. Designing with AI thus requires a prompt; without quality input, AI cannot turn dross into gold. Hence, a talented Hemanta-like singer is needed before AI can refine and convey nuances. This is where talent and effort diverge.

So, is this an exercise in sophisticated artistic engagement? Not at all. Yet, it signifies experimentation with novelty. In mimicking these voices, we might uncover something extraordinary. Could a new tune featuring Bilayet Khan's sitar in a modern Bengali song be considered a significant artistic achievement? Perhaps not. But what if AI helps create a mesmerizing composition? That would be noteworthy. It seems we should break free from all known conventions, whether we err or succeed, to try something new. At life's end, we can at least claim, "We tried." Technology is not our adversary but our ally. Despite all changes, Shreya Ghoshal remains Shreya Ghoshal, Arijit Singh remains Arijit Singh, and Zakir Hussain remains Zakir Hussain. The fear of job loss necessitates a different approach. Although computers have displaced many jobs, they've also created far more opportunities. Remaining insular is perilous. We must open every conceivable door to possibility. Let's then see what unfolds!

The next challenge: truth versus falsehood. Our civilization has battled counterfeits throughout history. Some create; others imitate for profit. Humans have deceived and will continue to deceive, from currency to music. Gandhi's visage on currency notes demands scrutiny for authenticity, often indistinguishable at times. Jamini Roy's paintings have been copied and sold for profit, hardly distinguishable to the layperson.

Fakes are ubiquitous: perfumes, watches, medicines—perhaps two trillion dollars worth of counterfeit goods are sold as genuine annually. Music, a product of capitalist society, is not immune to imitation. Once, piracy thrived; now, AI-generated fakes advance the issue. Yet, the music industry survived piracy and will adapt to this new challenge. Distinguishing AI-rendered Geeta Dutt from the original poses a new dilemma. Surely, solutions will emerge. Engaging with an AI Geeta Dutt is less a crime than spreading fake news. Trust in media is waning; news is scrutinized for authenticity. This malady is deeper, more dangerous. We aim not to become North Korea, do we?

Let AI present us with challenges. I welcome them. I aspire to create music far superior to my past work, with AI as my companion, my competitor. Let it counter my thoughts. We need our adversaries; without them, life loses its flavor.

Anupam Roy

(Opinions are my own.)