Art in the Age of Generative AI

There are more things in heaven and earth, AI, than are dreamt of in your data sets.

The spectre

It is about time that another of our manifestos began, ‘A spectre is haunting academics and the creative economy – the spectre of artificial intelligence.’

Various sentiments have been stirred by the arrival of generative artificial intelligence or AI in the domains devoted to intellectual output, and they are certainly more existential than mere disruption anxiety. 

The line of thought proceeds thus: If artificial intelligence comes, can artificial intelligentsia be far behind?

And what about artificial artists and artificial intellectuals? It turns out that the big war that is predicted to begin this year is possibly one of the most important in human history. Except that it will not be fought between countries, but will play out as a protracted tussle between the human mind and its artificial adversary. 

Unlike in the past, the impact of publicly available AI will be directly felt in cultural and social life, realms that have traditionally been less permeable to technological breakthroughs. As AI-powered by large language models, or foundation models, transgress into exclusive human preserves like language and art, we face a revisionist moment that needs to be discussed beyond technology and economics. 

The AI Pandora’s box contains a unique challenge for each walk of social and intellectual life: plagiarism in research, propaganda in the polity, tech-inequality in the economy and redundancy for the intellectuals and creatives. For all of known history, art has been a differentiator as well as a source of identity and pride for humans. What generative AI has taken away is the mystique and exclusivity of human creativity, raising the uncomfortable question: Are humans mere organic computers, un-upgradeable and limited by the slow pace of evolution, soon to be dwarfed by humanoids?  

The spectacle

As you go hands-on with the new AI kids on the block, the premonitions start to come true one by one.

I prompted AI to write a Shakespearean sonnet imitating Sonnet 18 ‘Shall I Compare Thee to a Summer’s Day?’ and suggested the title ‘Shall I Compare Thee to a Masala Dosa?’

 

Then it came up with the opening paragraph of a Sherlock Holmes mystery titled ‘The Case of the Missing Masala Dosa’.

A convincing portrait of Leonardo DiCaprio in the style of Raja Ravi Varma:

Realistic images of a unicorn in the streets of New Delhi:

The speculation

The internet made memory redundant by helping us externalise it onto servers and databases. Will AI do the same to decision making and creativity? As an economic question this may be answered away with the disruption theory: that the tipping point is when an alternative becomes good enough. AI is certainly good enough to disrupt commodified content and commercial art.

Their prestige counterparts, literature and academic art, are likely to survive owing to their organicity.

Products of human creativity are predicated on context and provenance, whereas machine minds only derive probabilistic outputs in the likeness of human art. Literature is produced, consumed and judged in constant reference to tradition, genre and the canon. It is these several layers of validation that assign aesthetic value. This is why an M.F. Hussain doodle earns millions while your own masterpiece struggles to get more than five Facebook likes!

At the moment, however, there are more questions than answers. Will we have a canon of machine art? Will humans and machines collaborate to produce hybrid or aided literature and art? How will the aesthetic value of these works be determined?

Let’s not forget that data is both the strength and the weakness of AI. On the one hand, it can pore through in an instant everything that has ever been written on any subject and create something new out of it. On the other, it can only go that far, and not further into that which has never been thought of or expressed.

As Hamlet might say, ‘There are more things in heaven and earth, AI, than are dreamt of in your data sets.’

That’s the only consolation for the moment.

Sreetilak Sambhanda is a publisher, editor and writer based in New Delhi.