How AI Will (And Won't) Change Creative Work

If you prompt ChatGPT to write a blog post, it will likely write you a pretty bad one. Not only will the copy come off as generic and derivative, but it may make up facts and citations.

It was widely reported earlier this year that the news outlet CNET issued corrections on several articles after using an AI-powered tool to write dozens of stories.

More recently, a lawyer used an AI-powered chatbot to create a motion that contained fabricated case law. The lawyer claimed he was unaware that the chatbot could provide inaccurate information, and is now facing possible sanctions.

Of course, there are examples of impressive feats generated by AI. "Heart on My Sleeve," a song that featured AI versions of Drake and The Weeknd, achieved viral success on social media, and enough accuracy to be removed for copyright infringement.

With the release of GPT4, it became possible to turn a picture of a napkin sketch into a fully functioning website. The technology also impressively scored in the top 10% in a simulated legal bar exam, while its predecessor GPT-3.5 scored in the bottom 10%.

AI has proven capable of impressive levels of competence, as well as mindless people-pleasing. It seems clear that for the foreseeable future, AI is a tool in a human-driven process, rather than the mastermind of it.

How AI Impacts The Creative Process

Producing creative works often begins by envisioning the end. A screenwriter writes in pursuit of a blockbuster movie; a company creates wireframes on a whiteboard, envisioning a popular website.

While the specific approach may vary across different fields, there is a traditional three-phase process for bringing creative ideas to life. At each phase in the process, AI can assist in the jobs to be done, thus lowering costs and potentially improving results.

Phase 1: Specification

The Specification phase is where the buyers or owners of a creative work provide goals and guidelines for creators. Usually, this is a budget, timeline, and requirements for how the end product should look and perform.

Much of the work involves research, benchmarking, and consolidating information necessary to produce the creative work. Inputs into product development, such as competitive analysis, customer research, style guidelines, and more, can be supported with AI.

AI can also play a role in bridging the divide between roles. In web design, product owners often struggle to communicate what they want, lacking the concepts and skills to produce meaningful modes of inspiration for their creators. This results in additional fiction in the web development process.

With AI, product owners can prompt and iterate through concepts, without the need for robust design skills, or expensive design sprints. When buyers can more clearly express what they need, creatives can work more effectively and efficiently within their vision.

Phase 2: Development

In the development phase, a creator or a team of creators, guided by a manager, translates the specifications into a product.

In this phase, AI helps apply common patterns and provides instructions for tricky engineering challenges.

For example, Webflow, a website publishing platform, and WordPress competitor, recently announced plans to tightly integrate their tool with AI.

Features include the ability to auto-generate Search Engine Optimization (SEO) settings based on the actual content of the page and create unique new images, components, and page sections based on a text prompt.

Many more tools are already all-in with AI. Notion, a popular document creation and project management tool, makes it easy to prompt AI to supply rough drafts, and templates, and to summarize and shorten text.

Phase 3: Publishing

The final phase in the creative process is publishing, when the creative work is released to the public, with hopes of reaching its intended audience and meeting expectations.

AI-powered automation can streamline time-consuming tasks such as content creation, ad targeting, and campaign optimization. Machine learning algorithms can optimize ad placements and budgets based on real-time performance data.

AI-powered A/B testing tools can quickly and accurately measure the impact of different product features on user behavior. These tools can provide product owners and creatives with insights into which features are most effective and which need improvement, allowing them to make data-driven decisions to further refine the product.

These features are being woven into some of the most popular analytics tools, including Google Analytics, which uses machine learning to help users better understand and act on data.

The Opportunity for Business

Creative work is vital to the success of any business. Products, and the marketing and selling of them, require imagination and creativity. And AI presents an enormous opportunity to build better, faster, and cheaper than ever before.

My prediction is that AI will empower smaller, more auteur-driven creative teams. Individuals that think about products holistically, and use AI to overcome their skill gaps, will work with a more narrow set of specialists — if any at all.

Are you ready for this future? Will you be able to compete against lean AI-supported creative processes? Now is the perfect time to start experimenting.