Hasta La Vista, Creativity?

It’s highly unlikely that one would be able to create an award-winning film, or a song entirely with AI.
Representative Image
Representative Image

Earlier this month, a slickly produced ad from a Silicon Valley tech giant, promoting its latest product – an exceedingly thin tablet – went viral for all the wrong reasons. The ad depicted musical instruments, art supplies, film equipment, and more being crushed under a hydraulic press, with the tablet emerging out of their destruction. Likely intended to demonstrate the product’s ability to do everything these creative tools could do, the ad instead sparked outrage among the creative community. Some took offence with the implication that a single device could substitute for the ‘human experience’ of a whole range of artistic tools. Others criticised the ad for its cultural insensitivity, given the historical and sentimental value many of these destroyed items hold in various cultures. For many, the ad was yet another example of Silicon Valley’s ‘callous attitude’ toward the creative community.

While not the first of its kind, the strong response to the ad seems to be largely due to its timing. Over the last 30-odd months, the relentless release of Artificial Intelligence (AI) tools, each touted as more capable than the last, has increasingly threatened to redefine what constitutes art and what it means to be an artiste.

Want to craft your own songs, complete with custom lyrics and genres? Udio’s got you covered. How about ultra-realistic artwork? Dall-E or Midjourney can conjure it up. Need to create a video but lack the resources or CGI expertise? No problem, Sora or Synthesia are at your service. And let’s not forget the large-language models behind the likes of ChatGPT, Claude and Gemini, which have largely dominated AI discourse. The sheer competency of these tools, despite their apparent infancy, has been so striking that major Hollywood studios have seriously considered using them to supplement, or even replace, human writers and artistes.

So, it’s no surprise that there’s an ever-growing sense of unease within the artistic community. As AI tools continue to evolve at a mind-bogglingly rapid pace, the (near) future of creativity appears increasingly dystopian. One where skills and human expressions are marginalised in favour of cold, efficient algorithms. But is the future really that bleak?

I’m no expert, but ever since 2015, when Google’s big-money acquisition, DeepMind, started quietly demonstrating the feasibility of programmes that can train themselves to do stuff that conventional wisdom until then suggested only humans could do, I’ve been down the rabbit hole, trying to understand the tech behind it – Machine Learning and Neural Networks. These concepts have been around for ages, but it’s only in the last decade that we’ve finally had computers powerful enough to put them to good use. But there’s still a huge problem that humans haven’t solved – nobody, not even the developers of these models seem to have a complete understanding of how they work.

This has led to companies behind these technologies having lesser than the desired control over how they behave. Chatbots spewing hate speech and image generators struggling with diversity are just the beginning. The fundamental problem with current AI models is their lack of originality. They are, at their core, mathematical models trained on existing data, merely imitating human output from the past. And hence, can never really replace human creativity – at least not anytime soon.

Meanwhile, as the single-biggest source of training data for these models – the internet – gets more and more saturated by AI generated content over the next few years, the quality of AI outputs is bound to decline. Unless, the models are fed with a steady stream of ‘fresh’ human generated data. How ironic!

The beauty of these models, however, mirrors the open-source software movement of the ’90s: they eliminate the need to start from scratch, providing a springboard for inspiration. It’s highly unlikely that one would be able to create an award-winning film, or a song entirely with AI. But you might be able to make money selling your ‘art’ to unsuspecting customers, as one Bengaluru teenager successfully showed last week!

(The writer’s views are personal)

Open Space

Dese Gowda



Related Stories

No stories found.

The New Indian Express