AI, writing and creativity

A mini-lecture

This post is part of a series which focuses on the role of generative AI in UK education. Starting September 2023, it aims to chart my journey through this rapidly changing area as I consider the practical implications of AI for both teachers and children. Please note, older posts may no longer be relevant. Delve in and please do get in touch with any thoughts.

Generative AI has thrown up enormous questions about copyright, intellectual property, and the future of creative industries. Last week, the IMF warned that AI may impact 40% of jobs and many artists already report reduced work opportunities as companies turn to AI image creators instead of illustrators for original artwork. And as of October 2023, ChatGPT4 now has integrated Dall-E 3 image creation, making it even easier and quicker to create images in a matter of seconds.

I’ve found that the children I teach are highly aware of these developments. I was recently asked to deliver a lecture for sixth-formers on AI and creativity and I was blown away by the level of interest and insight shown by the students. For anyone who might be delivering an assembly or lecture on a similar theme, my slides might be useful to you. This post is an outline of that lecture.

We begin by identifying which of the following was created by an AI. Play along (answers at the bottom)…

Which artwork was created by AI?

Which ‘Taylor Swift lyric’ was created by AI?

Which ‘Donald Trump quote’ was created by AI?

Which photograph was created by AI?

For most of us, it’s near impossible to guess which is created by humans and which by an AI.

But does it matter? Do we need to know from where our art and writing emerges? Do we need knowledge of the author? What happens if we don’t?

The death of the author

Some of these debates aren’t new. In the late 19th Century, a group of artists and writers decided that art was best divorced from morals, meanings and authorial purpose. “L’art pour l’art”, they said, asserting that artistic works exist to entertain and delight, not to moralise. These thinkers would consider modern authorship irrelevant: who cares if the creator is a robot or writer! What matters is how the creations are received and enjoyed.

Similarly, in the 1960s, French philosopher Roland Barthes argued that authorial intent is restricting. We should pay no heed to what an artist or writer intended to create; instead, all meaning and significance is in the eye of the beholder. Authority should be freed from the author; we should metaphorically kill off that writer and their limiting intentions. It is the reception that counts. Would Barthes be invested in knowing whether something was created by a human or AI? Possibly not.

But truth? Morals? Integrity? Accountability?

In our modern age, most people agree that some understanding of authorship is important. It’s right that we know if we’re experiencing art created by a human or AI. It’s vital that we have some sense of the contexts of creation. Here’s why:

1. Truth

In a world full of fake news and dubious claims, we can’t begin to scrutinise truth or reliability without a sense of who writes or creates. Critical literacy relies on context. And art is powerful. It doesn’t just entertain or delight; it also influences thoughts and inspires action. We need to know the author in order to discern the underlying motives and perspectives of information and art. When we understand the context and intentions behind a piece of content, we’re better equipped to evaluate its credibility. And because Generative AI masks it’s sources of information, it’s impossible to read it with a critical — and therefore fully comprehending — eye. As I’ve previously written, the only way to combat the dangers posed by new technologies is to throw everything at reading and digital literacy.

2. Understanding

Art grants us windows into different lives and perspectives. As Monica Ali writes, “we read to connect with human experience, human instincts and emotions”. Uncertainty in human authorship puts a barrier up to sharing experience and building empathy with others.

3. Accountability

If art is powerful, it’s crucial we remain attune to the biases that can influence its creation. This is true of both human-created art and AI created art and text, where LLMs are found to replicate significant gender and racial biases — they are, after all, trained on text created by our flawed society. Understanding authorship is vital in interrogating biases or inequities in artistic representation.

The benefits of AI to creativity

That’s not to say that there aren’t huge benefits to AI-created art and writing. Its speed, for one. I like to think that one day, AI can be used as a calculator for writing. We will still need to learn how to construct sentences and write independently, just as we still need to know basic arithmetic. But for the right tasks, perhaps we can use generative AI to leapfrog those more foundational processes of sentence construction that can take so much of our cognition, to free up more space for bigger ideas. Instead of labouring over clauses and commas, perhaps we can outsource these more operational parts of writing to spend more time on overarching argument, purpose, impact of our text. Instead of the brick-layers, it can help us be the architects of writing.

And AI has already seen new innovations in art. Notable examples include this astonishing prize-winning photograph and this amazing AI-created short film which criticises — perhaps ironically — the rise of AI in the creative industries.

And the drawbacks

But we’re right to stay cautious. Right now, we’re in the wild west of unregulated AI development. Many, many people have called for greater regulation. We’ve seen biases proliferate, such as the problematic mass reproduction of the “Afghan Girl” image. We’ve seen artists’ intellectual property be stripped. We’ve got ongoing court cases of copyright infringement as LLMs sweep up the internet with no consideration of the work gone in. We have the proliferation of chatbots, spewing hate designed to divide and incite. We have the many, many examples of misleading or completely inaccurate information caused by AIs hallucinating, valuing a quick response over accuracy.

For teachers, whilst maintaining a wide-angle view of the benefits and drawbacks technology poses, we also need to remember the purposes of creating art and text for learners. We don’t write just for a glossy, final piece. It’s not just about the outward-facing products we create. We write and create to process the world, to synthesise our knowledge and to represent our developing thinking in different forms. In short, creating art and writing isn’t merely a product of learning; it is a process of learning. And this just cannot be outsourced to technology.

Answers

Which artwork was created by AI? B

Which ‘Taylor Swift lyric’ was created by AI? A

Which ‘Donald Trump quote’ was created by AI? A

Which photograph was created by AI? A