On AI and the Problem with Instant Answers

Why the Humanities and the Arts Matter More Now Than Ever.

I like new technology. I am dictating this text to my phone, and I practice the art of bending ChatGPT to my will on a daily basis. The combination of GPT 4.0, Bing and Midjourney is deeply embedded in my workflows, and I cannot wait for the MS Office suite to launch with AI copilot functionality shortly.

At the same time, I find the transformative potential of AI one of the most disconcerting truisms of our time. We are repeatedly told that it is our uniquely human qualities that will shape how we coexist with AI in the future, yet these very qualities are already being overlooked in disturbing ways.

Of course, apprehensions about AI are not new. Elon Musk, Paul Christiano, Yuval Noah Harari, and many others have voiced concerns, and OpenAI CEO Sam Altman recently testified at a surprisingly informative Senate artificial intelligence hearing. But the backdrop for this post goes back a bit farther; Geoffrey Hinton, aptly named the Godfather of AI, recently made headlines for vocal criticism of AI after leaving Google, but the same Hinton delivered other salient points almost two years ago.

Before the AI frenzy became the new normal, Hinton poignantly called for a much broader involvement of the humanities and the social sciences in the development of new technology. He warned of the dangers in a technically advanced society if appropriate steps towards involvement of multiple disciplines, outside of tech, are not taken in a timely manner.

We are now witnessing manifestations of these concerns. In fact, there are already indications of a threat that is making us collectively more foolish and less capable as human beings.

The Chatbot Problem

Regular readers in this space may have seen The Problem With AI Chatbots in Literature, a post in which I discussed a TED Talk where Sal Khan of Khan Academy introduced his educational chatbot, Khanmigo. During his presentation, Khan related an instance where a young reader grappled with the symbolism of the green light in F. Scott Fitzgerald's “The Great Gatsby”. After having consulted a number of sources to no avail, she turned to Khanmigo and found herself engaged in a conversation with a digital simulacrum of Jay Gatsby himself.

Setting aside for a moment the degradation of the novel as an art form by suggesting a chatbot could somehow enrich a literary work, and ignoring the educational chatbot's flagrant disregard for centuries of literary history and criticism, the essential problem is this: “The Great Gatsby” is not in need of AI intervention or repair work. Its place in the world cannot be separated from its ambiguity, its power to evoke thought and imagination, its capacity to stimulate critical thinking, and a hundred other things we as humans bring into the experience of reading and, more broadly, interpretation of artistic creativity.

But these types of friction appear unwelcome in the age of AI, and I find this deeply unsettling. Literature, and more broadly speaking the humanities and the arts, rarely offer clear and unambiguous answers. And that's precisely where their value lies.

Art in the Age of AI

Another example is the many “text to xx” apps, such as text to video, text to image, text to music, etc. While it can admittedly be both fun and productive to use these apps, it is also deeply problematic. Almost a year ago, Jason Allen won the Colorado State Fair’s fine arts competition with “Théâtre D’opéra Spatial”, an AI-generated artwork. Since then, things have only gotten murkier in terms of misinformation, ethics, and copyright.

The broader implications for quality and originality are of pivotal human significance in this context, too. Historically speaking, a great deal of art has fundamentally refused the categories we so desperately want to ascribe to it, with words. However, an inherent dismissal of categorisation is hardly an ideal fit for generative AI: large language models prefers - well, models. And language. Does this mean we will now settle for less as human beings in a world where convenience seems to beat originality? While Allen did win the fine arts competition mentioned earlier, GPT 4 will not output Shakespearean masterpieces, and there is no new Picasso or Beethoven on the horizon either.

The Unreplicable Assets of Humanity

As we grow more reliant on AI to solve our problems and answer our questions, we also run the risk of mistaking the AI's creative abilities for our own. This could lead to overconfidence in our understanding and underestimation of our ignorance, a dangerous combination that could stifle our intellectual growth and artistic aspirations.

When we ask chatbots to explain complex concepts such as the theory of relativity in a way that would convince a cat to take a bath, or explain the rules of basketball using only emojis, we not only succumb to the convenience and amusement of AI but also risk falling into a pattern of not really trying. I wonder if we are already well underway in sacrificing our capacity for critical thinking and creative expression at the altar of instant gratification.

In this fast-paced world, with all its wonders and horrors, the allure of quick and easy answers is undeniable. But deep understanding and genuine learning often come from wrestling with complexity, from asking difficult questions and seeking impossible answers, from the challenge and the struggle. If we let AI take over these processes, we risk becoming spectators of humanity, rather than active participants in it.

It is astonishing to see how quickly AI has become second nature to us. It holds tremendous predictive powers for humanity, no doubt. But our humanity, our ability to think, question, and create, is our greatest asset.

No AI, however advanced, can replicate that.

Yet.