A lot of the noise about artificial intelligence and creativity is absurdly simplistic – far too many people are currently buying into the hype about creativity at the touch of a button.
The idea of instant creativity is one reason many creatives hate AI, as a parasite on their work and a threat to creative jobs. Not to mention their work being used without permission for AI training.
And for those of us who have spent years honing a craft, there’s a deeper anxiety: that the ‘efficiency’ of AI comes at the cost of the human spark. It feels less like a tool and more like a replacement for the very soul of what we do.
Which is a shame. Because AI does have a lot to offer in the creative sphere – just not in the way many techno-evangelists are claiming.
Professional artists and creatives know that real creativity is complex. It’s an iterative process, involving different types of work at different stages. For us, the real question isn’t ‘Can AI be creative?’ – it’s ‘How can AI amplify human creativity?’
AI and Human Creativity: It’s about Process, Not Just pressing a button
As a poet and a creative coach, I know the creative process as a practitioner, and also as a facilitator of others’ creativity.
It’s possible to identify several aspects or stages of the process, that are broadly similar across the arts and creative industries. And there’s no reason why creators can’t use AI at any or all of these stages, while retaining ownership of the vision and finished work.
If you’re reading this, I imagine you are a professional artist or creative, who may have reservations about aspects of AI, but who is also curious about what it can bring to your creative practice. I’m writing this to share what I’m learning, in the hope you will find it useful.
Before we dive in, here are a few assumptions I’m making:
- You are the creator; AI is an optional tool/collaborator/assistant.
- You are open-minded about AI’s potential, while sceptical of the hype and hysteria.
- You want to preserve and practice your core skills, and use AI to extend, not replace, them.
- You want to respect others’ intellectual property and protect your own.
I’ve gathered a lot of ideas and examples in this article, to give you practical ways to use AI in your creative process, in seven different areas:
- Developing Ideas
- Research
- Sketching and Drafting
- Doing the Previously Impossible
- Collaborating with Yourself
- Feedback and Revision
- Pitching and Promoting
1. Developing Ideas
Yes, AI can generate ideas, of varying quality. But coming up with new ideas isn’t generally a problem for creative professionals. We usually start with at least a kernel of an idea. For us, AI is more useful at helping us develop our ideas.
We typically do this either alone, with a notebook, sketchpad or voice recorder, or in dialogue with a trusted friend or collaborator. Now, AI gives us a third option at this early stage of the creative process…
Where AI can help
Starting a conversation with AI can be a great way to both capture and extend your ideas.
AI isn’t a substitute for talking to a human being, but it does have two advantages over most people you know: it’s always available, even at 3am, when an idea strikes and you want to talk about it; and it has instant deep knowledge across most domains, that can be invaluable in exploring the potential of your idea.
You can’t rely on it to get all the facts right, so later on, you’ll need to fact-check rigorously. But for this early stage, its overview of a topic is generally good enough to explore possibilities.
ABBA’s Björn Ulvaeus is currently incorporating AI into his creative process as he writes a new musical, as he told an audience at SXSW London:
A misconception is that AI can write a whole song. It’s lousy at that – very bad. And thank God! It’s very bad at lyrics as well. But it can give you ideas.
You have written a lyric about something, and you’re stuck maybe, and you want this song to be in a certain style. So you can prompt the lyric and the style you want, asking, ‘Where would you go from here?’… And it usually comes up with garbage, but sometimes there is something in it that gives you another idea. That’s how it works.
It’s like having another songwriter in the room with huge reference frames. It is really an extension of your mind.
The Hollywood Reporter
Things to try
Talk to the AI as if you were sharing an idea with a friend. Using voice mode makes it more like a real conversation, but voice vs text is a matter of personal preference. Either option gives you a transcript that can be useful at later stages of the project.
I have this idea… what do you think?
Here’s my idea in three sentences. What parts feel strongest, and what feels generic or overfamiliar?
As the conversation develops, ask it to help you extend the idea:
What else can I add to it?
What else could this be?
Who would value it most?
Where else could it be useful?
Writer and founder Stew Fortier asks AI to ‘reverse interview’ him, by asking questions about the idea. Here’s a prompt (not Fortier’s) to get started:
Act as a curious interviewer. Ask me ten questions that would help you understand what’s really interesting about this idea.
You can also get it to nudge you into action:
Where should I start?
What’s the minimum viable version of this idea, that I could show to potential collaborators?
If I were on a shoestring budget, what version(s) of this idea could still work?
Boundaries and pitfalls
Don’t rely on AI to do all the heavy lifting for you. Research has shown that group brainstorming sessions work better if everyone has done some work beforehand, generating ideas on their own.
My experience is that the best AI sessions are often like this – I work on the idea myself in my notebook, then talk to the AI once I’ve reached a natural pause, and I want another perspective on it. This makes the AI conversation more productive – it gives the AI more to work with, and crucially, it means I’m developing my ideas rather than the bot’s.
2. Research
Ideas need to be grounded in reality if they are to work, most obviously in the case of non-fiction books, or the laws of physics for buildings or furniture.
Even if you’re writing a fantasy novel, you want to make sure your storyworld won’t seem too familiar to aficionados of the genre. And you definitely want to avoid plagiarism. When the melody for ‘Yesterday’ came to Paul McCartney in a dream, he worried that he might have heard it before:
For about a month I went round to people in the music business and asked them whether they had ever heard it before. Eventually it became like handing something in to the police. I thought if no one claimed it after a few weeks then I could have it.
But that was yesterday. Today’s creators can save themselves a lot of time in the early stages of a project, by enlisting AI as a virtual research assistant.
Where AI can help
For those of us without a large research library on our doorstep, AI research tools are a big step forward compared to searching Google, article databases, library catalogues etc. AI can search and synthesise much faster than a human, so instead of a collection of links, AI can give you an overview of a topic, with recommendations for key texts to read.
You can also ask AI to help you test your assumptions, and look for evidence for and against your hunch or hypothesis. In a commercial context, market research can be crucial.
AI can help you avoid the kind of unconscious plagiarism McCartney was concerned about. More than that, it can be a great ‘originality filter’ – give it your idea or a sample of your work and ask it how fresh and original it is, compared to what has been done before.
Some artists are even using research data itself as material for new kinds of art, such as the stunning ‘Archive Dreaming – AI Data Sculpture’ Refik Anadol created for SALT Research
Things to try
To get started on a new topic, ask:
Give me an overview of this subject, including where the consensus is strongest, and what issues are still contested.
Who are the leading experts on this topic? What do they say?
Give me 5 essential books to read on this topic, that cover it from different angles.
Test a hypothesis using academic AI research tools, such as Scite, Elicit or Consensus:
My hypothesis is X… but what does the research say?
What supporting evidence can you find? What challenges it?
Cite sources – and tell me what each source supports.
Avoid inadvertent plagiarism using tools such as Scribbr (writing), MIPPIA (music), and AI-based visual similarity search tools (e.g. Google Lens).
Discover how original your work is, by asking an AI:
Has anyone done something meaningfully similar before?
And direct the AI to search in specific areas, for example:
Writers: search across genres, not just your own. Look for:
- earlier novels, short stories, essays
- mythological or folkloric precedents
- non-fiction treatments of the same theme
Visual artists and designers:
- Museum and gallery archives
- Reverse image search
- Design history books and patent databases (for functional objects)
Filmmakers and media creators:
- Festival catalogues
- Interviews and production notes
- Earlier works by the same director or movement
AI will almost certainly find some similarities with previous works – it’s your job to decide whether the overlap is superficial, problematic or complementary.
Boundaries and pitfalls
Large Language Models (LLMs) are notorious for hallucinating, so always ask for sources! I’ve trained ChatGPT to give me a literature overview and confidence rating (‘low-moderate-high’) about its findings on any given topic.
And when the outcome really matters, and you’re dealing with a subject outside your field of knowledge, there comes a point where you need to talk to an actual human expert or researcher, who can give you the in-depth knowledge and reassurance you require. Because if something goes wrong, the buck will stop with you.
3. Sketching and Drafting
This is where the talking and thinking stops, and the actual creating begins: the writer’s first draft, the artist’s sketches, the screenwriter’s storyboard, the musician’s demos.
For many creators this is the most exciting stage of the creative process, when ideas start to become real and concrete. AI offers us some intriguing new options at this stage – and also some important decisions.
Where AI can help
Here’s where I’m a little hardcore: as a writer, I insist on writing the first draft myself. Especially when it comes to poetry. The idea of asking a bot to write the draft of a poem strikes me as ridiculous. Even when it comes to prose, I write the draft myself. Otherwise I feel like I would be missing out on something essential.
But I’m at one end of a spectrum here… I know some excellent writers who have trained an AI on their entire back catalogue and use it to generate a first draft, which they then edit extensively, to get to a version they are happy with.
Many writers, screenwriters and directors use AI for structure and scaffolding – outlining chapters, or building storyboards, beat sheets and shot lists, which are then worked over by humans.
Another emerging use of Generative AI is to create concept art and other visuals to aid discussion and collaboration between professionals with different skill sets. A character or scene visualisation can help a screenwriter convey the essence of a script to a director or cinematographer.
Zaha Hadid Architects use AI in client presentations, to render design options during a meeting, so that ‘it helps [clients] participate in their own building’.
Things to try
Generate multiple rough versions of the same thing and develop the one(s) you like:
Give me 5 radically different openings / thumbnails / 8-bar loops / blocking ideas for this concept. Keep the core premise, change the energy and form.
Ask AI to help you develop a structure for your work, such as the outline of your book, a storyboard, beat sheet, etc.
Create collaboration material – such as character and scene visualisations, diagrams, flow charts, and promo materials. Crucially, these don’t need to be finished to a professional standard; they are to facilitate discussion and collaboration, so as long as they spark a useful conversation, they are doing their job.
If you’re feeling brave, and if you trust the model’s security settings, you could upload your work and ask it to help you generate new drafts and sketches in your own style.
Get more original or edgier results with an anti-blandness rule – whenever you ask AI for options, also ask for one version that deliberately violates your own habits or preferences, and one that risks being unpopular or uncomfortable.
Performers – generate rehearsal prompts:
Create 8 improv prompts / movement scores / acting objectives for this piece, each with a clear playable constraint and a 60–120 second duration.
Boundaries and pitfalls
The more material you allow the AI to generate on your behalf, the bigger the risk of blandness, or of losing touch with your craft skills. You also risk freezing your style – if the AI is creating options based on your current style, you are less likely to evolve into a new one.
And however thoughtful and skilful your use of AI, using it at all can be controversial. Pentagram attracted heated criticism for using Midjourney to create a set of 1,500 logos for a US Government website – even though they started the process with their own handmade paint and cut-paper designs.
4. Doing the Previously Impossible
Artists have always used new technology to extend the possibilities of their medium.
The ancient Greeks used stage machinery to enable gods to descend onto the stage as if from the heavens.
In the 1920s, the development of high quality microphones was exploited by the crooners – singers who used the intimacy of the microphone to create a new type of seductive vocal performance.
In just over a century, filmmaking went from blurry black and white silent pictures, to immersive HD worlds that could never exist, with actors playing characters who look completely different to them, such as Andy Serkis’ Gollum.
AI once again gives creators the opportunity to radically reinvent and extend their work.
Where AI can help
In ‘Creativity and Artificial Intelligence’, an influential paper published in 1998, the cognitive scientist Margaret Boden proposed that ‘AI techniques can be used to create new ideas… by making transformations that enable the generation of previously impossible ideas’.
Now we are entering a period of experimentation, as artists and creatives explore the potential of new AI tools.
Some screenwriters are going beyond using AI to create visuals for collaborating and pitching, and using the tools to create the actual movie. AI short films are becoming a new category, represented at the Reply AI Film Festival.
While many artists are concerned about their work or image being used by AI, others are actively encouraging the idea. The musician Grimes wrote on X in April 2024:
I’ll split 50% royalties on any successful AI generated song that uses my voice. Same deal as I would with any artist i collab with … Feel free to use my voice without penalty.
The artist Sougwen Chung sees AI as an opportunity to create a new kind of art, as described in an Observer article about her collaboration with an AI-powered robot in front of a live audience:
The artist begins planting broad brushstrokes onto the canvas while wearing an EEG headset that monitors their brainwaves. Minutes later, they give paintbrushes to the bespoke robotic system… which has been fed with Chung’s movement and spatial data that lets it paint on its own. Working in unison, they and it engage with each other without using words; instead, their conversation is recorded on canvas in shades of blue.
Vibe coding removes technical barriers to entry when it comes to building software: non-coders tell the AI what they want to create, and ask it to write the code for them. Like AI films, vibe coding is spawning new genres, via app hackathons and game jams.
Things to try
The simplest way to get started is to ask the AI to be your guide:
I’m a [writer/artist/musician etc.] Here are some examples of my work.
What new creative possibilities can you offer me?
What can you help me do, that I couldn’t do before?
Or think of something you’ve always wanted to do, but assumed it was out of your reach – because you didn’t have the skills, collaborators, budget, or other requirements. Tell the AI about your dream project and ask it to help you.
Experiment with a new medium. If you’re a writer, trying making music. If you’re a musician, try painting or movie making. And so on.
Try vibe coding. If you could create an app or a game or another piece of software, what would you build? Describe it to the AI and ask it to help you build it.
Treat all of the above as experiments and don’t get too perfectionistic about them. You may not end up with work that is finished to a professional standard, but playing with AI can help you stretch your imagination in different directions.
Some experiments will be more successful than others – you may be surprised to discover how easy your idea is, or how hard. But don’t worry about this in advance. Tell the AI what you want to do, and ask it to act as your guide and lead you through the process of making it happen.
Boundaries and pitfalls
Attempts to work in unfamiliar media often expose the fact that technical skills aren’t the only barriers to entry. A writer might create a film that looks superficially impressive, but because they don’t have the skills of a director, cinematographer or film editor, it drags for the viewer.
Vibe coding has a lot of exciting potential – and often faults or security vulnerabilities that limit its usefulness.
Even when an artist overcomes the artistic and technical obstacles and produces work that finds an enthusiastic audience, they can still encounter legal issues and industry backlash.
In 2022 Kris Kashtanova wrote a comic book, Zarya of the Dawn, and illustrated it by generating images with Midjourney. But the US Copyright Office revoked protection for the images, after concluding they were not human-generated. And many in the comics industry have taken a strong stance against AI-generated art, so creators like Kashtanova risk the ire of human illustrators.
So if you want to take your AI experiments out of the studio and into the marketplace, you may need a thick skin for criticism and a good copyright lawyer.
5. Collaborating with Yourself
Many artists have built on their own previous work, incorporating themes and motifs from their previous work and developing them in new ways. Think Rembrandt’s self portraits, Yayoi Kusama’s dots, and Honoré de Balzac’s La Comédie Humaine, with characters reappearing across different novels.
AI gives creators new ways of exploring and experiencing their own work, and new ways to ‘collaborate’ with their own past selves.
Where AI can help
Whereas Doing the Previously Impossible is about breadth, extending yourself outside your usual creative domain, Collaborating with Yourself is about depth – making new discoveries in the latent potential of your own oeuvre.
I use the word ‘collaborate’ because the user-friendliness of today’s AIs means we can engage with our own work in playful, insightful and potentially creative ways.
In his book The Creativity Code Marcus du Sautoy describes The Continuator, an AI musical jamming partner, trained on a musician’s own work, and quotes jazz musician Bernard Lubat:
the system shows me ideas I could have developed, but that would have taken me years to actually develop.
It is years ahead of me, yet everything it plays is unquestionably me.
Watch a video of Lubat jamming with the Continuator.
The DJ and producer Reinier Zonneveld extends the jamming from the studio into performance, playing live sets in tandem with an AI clone of his ‘musical brain’:
To my knowledge, no one else has tried a project like this, partly because I don’t think anyone has access to as much data as we did… I’m such a music nerd, I have all of my recordings… which has put me in this unique position of being able to train AI purely through my music.
Music Radar
Choreographer Wayne McGregor has collaborated with Google Arts and Culture Lab on AISOMA, a choreographic tool fed on hours of footage from McGregor’s archive.
It generates new options for movements, based on the archive footage, which McGregor can incorporate if he likes them:
The choices are still mine. I’m the source, I’m the person at the beginning,” he says. “I see this more as an opportunity – I love being in the game.
Wired.
Painter David Salle has trained a custom model on his previous work to generate starting points that he then transforms into finished paintings. Far from freezing his style, he says machine learning allows him to explore many more possibilities than he would have time for on his own:
The machine can synthesise things in a matter of seconds… This evolution in painterly terms might take years or even decades.
The Guardian
Things to try
AI can help you analyse your own work in different ways.
A fun place to start is to upload a piece of work to Google Notebook LM and ask it to perform its party trick of generating an Audio Overview – a podcast episode with two (virtual) presenters discussing your work.
Other ways you can use AI to analyse your work include:
- Theme mining: ‘List the recurring obsessions in my writing/painting/music; show me supporting passages.’
- Motif map: ‘Generate a mind-map of phrases/images//melodic patterns I keep returning to.’
- Identifying polarities: ‘What tensions or contradictions are central to my work?’
- Blind spot spotter: ‘What’s missing from my work? What could I be avoiding?’
- Continuity checker: for long projects, have it build a ‘world bible’ (characters, rules, timeline) and flag contradictions.
If you want to go beyond analysis, you can upload your work to an AI (caveats below) and ask it to generate new options based on your previous work – images, plot ideas, melodies, physical movements, etc. You may get better results by giving specific prompts based on your own knowledge of the crucial elements at play – e.g. ask it to keep one element the same (palette, tempo, character desire, movement quality) while changing another element.
Or you might like to try artificial synaesthesia – upload a poem and ask the AI to turn it into music. Or turn your music into a painting, or a video, or a poem, etc.
Boundaries and pitfalls
Be very careful about uploading your work to any AI. Read the small print and check the security options – especially whether or not you are granting permission to use your work for training.
Personally, I’m OK with uploading the books for creatives that I’ve already published, but there’s no way I would upload the unpublished manuscript of my poetry collection.
An obvious creative risk is becoming too dependent on the AI for new ideas, and cannibalising your own work, so that it feels imitative. So it makes sense to also keep developing completely new work in your usual way.
6. Feedback and Revision
This is the ‘good to great’ of creative work. In spite of the popular image of inspiration, very few great artworks have been channeled in a single burst of creativity, and presented to the world without significant reworking.
Often, the difference between pretty good and apparent genius is a willingness to keep coming back to the work, looking at it from all angles, and refining it. AI can give you a lot of new options in this area, and can even make this part of the process enjoyable.
Where AI can help
For me, this is one of the most valuable uses of AI. I like to write a draft and then ask it for different types of feedback: starting with logic and fact checking, and moving on to the more imaginative process of asking it to adopt the persona of my ideal reader(s) and tell me how they respond to my writing – emotionally as well as intellectually.
I haven’t completely rewritten a piece based on AI feedback; what typically happens is it shows me areas where I may not have fully taken into account a reader’s perspective or situation, which feeds into my thinking about the next draft.
I don’t use AI to write poetry, but when I showed it a few of my poems and asked it to describe my poetic style, it was surprisingly accurate and insightful. It also made a few connections between my performances on my poetry podcast and the style of my written poems that I hadn’t noticed before.
The same approach works for images, sound, performance and design – you’re using AI as a second set of eyes/ears.
There is an emerging market in custom-built feedback tools. Logo Rank is designed to critique an uploaded logo on criteria such as uniqueness, legibility, and color/contrast, then provide suggestions for improvement.
Photographer Trey Ratcliff has built a ‘virtual art director’ version of himself to critique photos.He trained a custom GPT on his own books and blog posts, to give compositional and aesthetic critique of users’ photos – essentially ‘How would Trey see this?’
AI can also provide technical help with post-production. If you are a photographer, filmmaker, musician, producer, film editor or another role that combines creativity and technology, there is a mind-boggling array of AI tools to take the grind out of the process while giving you control of the creatively crucial decisions.
Things to try
AI can help you improve your work with different types of feedback and a range of technical tools.
If you regularly work on the same kind of projects, create a ‘feedback flowchart’, to ensure that you evaluate each piece of work in a consistent and rigorous process.
Big picture patterns. For instance:
- Storytelling – narrative and emotional arcs, character development, inevitability vs. surprise, stakes and consequence, unity of theme
- Art and illustration – visual hierarchy, overall balance / imbalance, dominant vs. subordinate elements, harmony vs. tension, where attention rests vs. travels
- Music – overall emotional journey, tension and release, contrast vs. continuity, unity of sonic world
- Performance – attention arc, energy modulation, presence and authority, relationships (solo – ensemble – audience), contrast and surprise
Audience personas:
- Describe your ideal reader/viewer/listener
- Ask the AI to give feedback as this persona – How do they feel about the work?
- What do they like/dislike or agree/disagree with?
- Repeat with as many personas as you like – useful when your audience is composed of different types of people, and you can expect a range of responses.
An informed critique:
- Ask the AI to act as an knowledgeable critic, an expert in your field, and give you feedback based on their domain knowledge and professional standards.
- This may well differ from the audience persona feedback above, and can help you take your work to a higher level.
- Depending on your goals (and emotional sensitivity!) you can ask it to take a more ‘incisive’ or ‘encouraging’ approach.
- You can also ask for a critique from a certain aesthetic or intellectual viewpoint, if relevant to your work.
Technical refinement:
- Photography – culling, exposure and white balance, noise reduction and sharpening, retouching and cleanup, batch export, metadata
- Film and video – transcription and text-based editing, stabilisation, shot matching, rotoscoping and masking, captions and subtitles
- Music – noise cleanup and restoration, stems and source separation, timing and pitch correction, mixing, mastering
- Design – design system consistency, accessibility checks, layout and responsive checks, asset export, design handoff
Checking crucial details. Depending on your media, these can include:
- Fact checking
- Logical errors
- Spelling and grammar
- Copyright infringement
Boundaries and pitfalls
When it comes to making art to connect with humans, there’s no substitute for the real thing. Virtual beta readers can help you anticipate human response, they are quick and easy to use. But at some point you need to put it in front of the people you’re creating it for. There are no guarantees they will like it, but AI feedback can help you give them your best work.
And with just about every AI tool, beware of the tendency towards blandness and conventional taste; if necessary, ask for feedback based on your very specific taste and attitude, and give it examples of the kind of work you want to benchmark yourself against.
7. Pitching and Promoting
Even when your work is finished, it’s not finished. You need to present it to the world – by pitching potential clients or collaborators, or promoting it to your audience or customers.
AI gives you some intriguing new options in this sphere, starting with the emotional challenges.
Where AI can help
Creators often struggle with self-promotion. Sometimes this is shyness. But it’s also about being so immersed in the work that it’s hard to see it as others do.
AI can shift your perspective, bypassing your ego, and showing you the work as it is, and why it will (or won’t!) appeal to others.
It can also analyse a complex product – such as a book, movie script, or product design – and give you a structure that highlights the most attractive elements.
When I uploaded one of my books for creatives to ChatGPT and challenged it to write a better sales page than the one I had written, it gave me a genuinely fresh look at my own book. It highlighted things that were important to my ideal reader, but which I had left out. I rewrote most of the actual copy, but it was still a valuable exercise.
AI can also help you sprinkle some ‘previously impossible’ magic on crucial pitches. Business Insider shares a story from Elena Knox, when she was Executive Creative Director at BBDO New York:
Knox said she was using AI to sell ambitious concepts to clients. She and her team used tools such as Midjourney to bring to life a vision of an office where tree growth exploded into it before landing the viewer in a forest in the form of a GIF. It convinced the client to invest a multimillion-dollar production budget for shoots in New Zealand and Bulgaria.
‘Before you would have had to spend hours comping that, and the image wouldn’t have existed,’ Knox said. ‘In AI, we were able to make a quick GIF of it and show the clients: This is what we want to film.’
Things to try
Avatars
First, develop a clear image of your ideal reader, customer, viewer or user – what marketers call an avatar.
This is what I know about my ideal reader/customer/viewer/user. What else can you infer about them, that will help us clarify who they are and what they want?
Once the avatar is clear, ask:
Here’s my work – which aspects of this will be most attractive or useful to them?
Angles
An angle is a specific, truthful way of describing your work highlights one aspect of it that will be attractive to someone else.
Show your work to to the AI and ask it to list 10-20 angles, explaining its appeal to different people for different reasons. Then pick the ones that feel most compelling.
Types of angle to ask for:
- Craft / process angle
What did I do that was unusually hard, meticulous, or revealing? - Problem solved / value delivered
What does this work do for the audience – emotionally, practically, socially? - Origin story
What question, wound or fascination generated it? - Cultural moment angle
Why now? What’s happening in the world that makes this resonate? - Audience identity angle
‘For people who…’ - Comparative angle
Two or three comparable works – but with why (e.g., ‘X’s intimacy + Y’s structural weirdness,’ not simply ‘for fans of X’ only)
Authentic outreach
Used wisely, AI can help you find collaborators for your project, such as curators, producers, investors, editors and agents.
You can show it your work and ask for a shortlist. You can ask it to research each potential partner. You can even ask it to draft an outreach email, based on what it knows about you and them.
But you should never use any of this as a substitute for getting to know someone’s work, and giving serious thought to whether your project might be a good fit for them. When it comes to the email, the final draft should be authentically yours. Anything less is disrespectful to them, and a big risk for you.
Promo materials
The same goes for promo materials such as:
- Sales pages
- Pitch decks
- Press releases
- Proposals for clients or grants
- Emails to potential collaborators
- Summaries and show notes – e.g. for an hour long podcast
- Social media posts
It’s a risk to just ask the bot to write them for you. Treat its outputs as scaffolding or primers, to be revised into your own voice.
Boundaries and pitfalls
These days, everyone is on high alert for anything that looks AI-generated – so don’t believe the hype about marketing materials at the touch of a button. As with any creative work, AI is best used to help you iterate to a final product – or to feed into your brief for a marketing professional.
Final Thoughts
AI is not creative in its own right. But in the hands of a skilled professional, it has a range of applications at different stages of the creative process.
We don’t have to look far to see the pitfalls of AI, or examples of it being used badly. But used well, it can widen the field of possibility, amplify your thinking and reflect your own work back to you in unfamiliar ways.
Most of the examples I’ve given show creators using AI as a stimulus to their own thinking, scaffolding to be painted or drafted over, or a mirror to see themselves and their work in a different light.
However much (or little) you use AI, your taste and judgment are essential to producing great work.
I hope this piece has given you some food for thought – and your experiments bring some welcome surprises.
