For creatives, inspiration is fleeting. The fear of the blank page leaves projects stalled and ideas left unexplored. AI image generation models like DALL-E provide intriguing sparks — but tapping their potential manually has limitations. By combining these tools with automation though, we can unlock new depths of creative possibility.
The Rise of Creative AI
Recent advances in machine learning have led to exponential progress in AI capabilities. Models can now generate strikingly novel images, text, code, designs, music and more from simple text prompts. They build on massive datasets to learn high-level concepts and relationships. The results exhibit sensibilities far beyond rigid algorithms of old.
Most generation systems today utilize a method called diffusion. This starts with random noise and gradually improves the output through thousands of tiny refinements while staying anchored to the text prompt. It’s like developing a photo in a darkroom, revealing more clarity over time. The layers of abstractions these models have learned enable extraordinary feats of imagination and non-linear idea combination.
DALL-E 2 is one of the leading image generators. Created by AI research lab Anthropic, it can renders 3D-like depictions of concepts with lighting, shadows and compelling compositions. The startup Chroma builds on similar principles for video generation. Jasper steers towards more whimsical cartoon aesthetics powered by Google. Most continue the rapid pace of innovation seen since DALL-E’s original debut in early 2021.
Adoption of these new co-creativity tools is accelerating — 81% of those surveyed intend to use them over the next year according to Adobe. Extensions even let you generate images inside Photoshop. While debates persist on ethics and job impacts, assisted ideation today demonstrates immense untapped potential.
The Science of Inspiration
What makes these AI systems so surprisingly inventive? Insights from cognitive science and creativity theory help explain the phenomenal fluidity.
Much inspiration arises through novel combinations of existing concepts in what‘s called bisociation. We relate previously unassociated inputs into something fresh and valuable. Mental shortcuts help us recall relevant bits to connect while suspending judgment long enough to explore unusual ideas.
Generative AI models exhibit similar combinatorial novelty…scaled to an extreme degree. Their vocabulary spans an unfathomably large conceptual network reflecting statistical regularities in massive datasets. This allows prompting new linkages we likely wouldn‘t consciously conceive of. The strange, counterintuitive associations often end up quite evocative and engaging.
Bisociative brainstorming with these tools hugely amplifies our capacity for serendipitous discoveries. We input some keywords and instantly stimulate unforeseen connections to build upon. It becomes easier to bypass creative blocks from habitual thinking and expectations. The unpredictable suggestions break us out of ruts.
While much art technology to date focused on efficiency and replication, AI finally brings ideative augmentation. We rapidly explore highly divergent concepts filtered through simulated artistic insight far beyond methodical human output.
Overcoming Bottlenecks with Automation
Manually using DALL-E proves enjoyable but inefficient over time. Each prompt requires contemplating what new idea to try generating. And it‘s easy to forget querying regularly enough to make it a consistent habit.
This is where combining the creative riches of AI with workflow automation provides game-changing impact.
Services like Zapier allow easily connecting apps into triggered sequences. We can setup generation to run automatically on schedules or events and route the outputs to any destination.
By integrating DALL-E and other creative tools into an automated pipeline, we solve several bottlenecks:
Prompt Inspiration
Instead of racking our brains to conceive ideas, we randomly pull inspiration keywords from a database. Unique combinations get constructed algorithmically every run.
Forgetfulness
Tuning into novelty exploration becomes a habit embedded into our regular workflows without manual effort.
Visibility
Outputs get delivered on a timeline we actually check instead of forgotten in an AI sandbox.
Flexibility
Mixing tools and remixing results opens exponential possibilities not viable manually.
Constructing these assemblies liberates the creative process from the drudgery of constant decision-making and unlocks vastly more novelty with no additional labor.
Step 1: Building a Prompt Database
The first component is a database holding all the keywords and concepts to mix into prompts algorithmically. This provides the raw materials for bisociative surprise.
We can structure this simply as a spreadsheet with columns for relevant prompt dimensions like:
- Nouns
- Adjectives
- Styles
- Themes
- Composition
Populate each column with diverse words and phrases that you find interesting or relevant to your domains of curiosity and expertise:
Tools like Dallery‘s Prompt Builder offer extensive ideas to draw from in compiling your database.
Ideally, include a wide breadth of vocabulary ranging from literal to abstract so the combinations induce thought-provoking interpretations by the AI.
Use duplications and weightings strategically if you notice certain words yield more intriguing results. For example, emphasizing "surrealism" and "cyberpunk" led to particularly unique suggestions in my testing.
Tweak and augment this vocabulary bank over time as you discover new concepts and refine interests.
Step 2: Automated Randomized Generation
With the building blocks assembled, it’s time to automate prompt construction and AI generation.
We‘ll use Google Sheets and Zapier to randomly pull words from our database and combine them into prompts formatted for DALL-E and other services.
Configuring the Google Sheet
Create two sheets named "Generated Prompts" and "Keyword DB". The latter will be populated per the previous step.
In the "Generated Prompts" sheet, use this formula in cell A2:
=CONCATENATE(INDEX(FILTER(Keywords!A2:A, Keywords!A2:A<>""), RANDBETWEEN(1,COUNTA(Keywords!A2:A))), ", ", INDEX(FILTER(Keywords!B2:B, Keywords!B2:B<>""), RANDBETWEEN(1,COUNTA(Keywords!B2:B))), ", ", INDEX(FILTER(Keywords!C2:C, Keywords!C2:C<>""), RANDBETWEEN(1,COUNTA(Keywords!C2:C))))This will combine a random selection from each database column into a new prompt every edit.
Adjust the range references and columns as needed to match your structure. Additional CONCATENATE add-ons easily incorporate more columns.
Now we have an endless supply of unique conceptual mashups to query our creative AI tools with!
Building the Automation Workflow
We‘ll construct the automation workflow in Zapier to trigger generative AI image and video generation from our freshly minted prompts.
The workflow structure consists of:
- Scheduler – Runs the process on a timeline
- Google Sheets – Gets latest randomized prompt
- DALL-E & Generative AI – Generate output from prompt
- Email – Sends creation to your inbox
Configure Schedule to execute daily or whatever cadence suits your needs. More frequent runs produce higher volumes of ideas to stimulate possibilities.
The Google Sheets step is set to update the randomized prompt cell, capturing the latest bisociative mashup.
Pass this prompt to DALL-E, Chroma and other AI tools via their Zapier integrations to output images and video clips. Most offer free tiers to start.
Finally, route the computer-generated creativity straight to your email or other destinations through Zapier‘s connectors. Some models support requesting variations on results as well for even more stimulus.
Watch your inbox fill up automatically with AI‘s imaginative suggestions daily, no manual intervention needed!
Maximizing Impact Across Domains
While the examples so far focus on artistic inspiration, the framework proves highly adaptable. The same approach applies to nearly any creative pursuit.
Some examples:
Writing – Populate words relevant to plot, character and setting elements. Automate prompts for scenes, dialogues and poems.
Design – Create mood boards, color palettes, icon concepts. Help envision UX flows, app prototypes, architectural sketches.
Music – Seed instruments, genres, song topics. Explore rhythmic foundations and melodic motifs AI composes.
Research – Brainstorm hypotheses, construct concept maps to identify connections between findings.
Business – Strategize positioning, branding angles and naming ideas. Envision user journeys and future interface possibilities.
The workflow essentially functions as an automated research assistant, providing external stimulation exactly when you need it. The key is curating the database vocabulary to reliably yield intriguing, on-theme suggestions.
While AI cannot fully replace human creativity, it excels as a launch pad for further refinement. The computer handles the mental heavy-lifting of combining concepts so we can focus on discernment and shepherding the most promising notions.
Advanced Integrations and Customizations
The previous guide covers the baseline workflow for getting started with daily AI-fueled inspiration. Many expansions and enhancements build on these fundamentals…
Some examples:
Chaining Multiple AI Tools – Route outputs from one model into the next for successive refinements and permutations. Remix combinations open up vastly more possibility spaces through compositing media types into collage art and videos.
Collaborative Innovation – Share sheet access to let teammates contribute ideas or help curate results. Automations distribute generated concepts across groups like writers workshopping scene prompts together.
Hybridizing Digital + Traditional Art – Print outputs to incorporate into hand-made works with painting, drawing or sculpture. Expands expression through blending AI‘s novelty with human touch and style.
Direct System Connectivity – Using developer APIs from tools like DALL-E, Stable Diffusion and Midjourney allows tighter integration capabilities. Build custom interfaces and functionality rather than rely on pre-packaged connectors.
Iterative Development – Create cycles of versioning advancements by feeding outputs back into the system as new inputs. The AI refines its own contributions through recursive loops.
Commercialization – Generate designs, merchandise and decor possibilities to test demand before manual creation. Automate prototyping of new products.
Outcome Analysis – Track performance metrics on prompts and ideas. Identify patterns around most successful keyword combinations, sentiment responses and more. Lets you empirically optimize the process over time.
The list goes on. Once the foundations are laid, we unlock nearly limitless applications for automated generative creativity in our workflows.
Democratizing Inspiration with AI
What seemed like far-fetched science fiction only a few years ago now feels commonplace. AI broadens creative potential for all by handling much of the diligence allowing us to focus efforts higher up the value chain.
Democratized access means no longer needing specialized skills or expensive tools to explore inventive ideas and make them tangible. Generative foundations let anyone rapidly build prototypes to evaluate and refine before committing more intensive resources.
For established experts and institutions, AI looks poised to fundamentally evolve creative processes as we know them…
These technologies keep advancing the frontier of possibility and inspiration. What we imagine today through imperfect tools represent just the earliest glimpse of a future aided by intelligence far beyond our own.
Sparking Your Creative Engine
Hopefully this guide provided some inspiring ideas to consider. Automating the repetitive inspirational heavy-lifting allows more energy spent crafting and connecting.
The core principles apply nearly anywhere novelty and invention play a role:
- Curate building blocks reflective of your interests
- Automate random prompt generation
- Trigger creation from AI‘s imaginative reservoir
- Review outputs and refine selections
Setting this cycle in motion requires only minutes but unlocks transformative creative leverage.
Now it‘s your turn. Try crafting an inspiration engine tailored to your own expression and domains. See where algorithmically-assembled randomness fused with AI‘s approximations of insight take you!
