Rather than fear AI, writers should learn how to use them properly. While this tech is transforming many sectors, and creative writing is no exception, it boils down to how unique a written content.
To this end, the Microsoft research team joined hands with the University of Southern California to experiment on whether generative AI boosts or weakens a writer's uniqueness.
Writers Meet AI: An Experiment
The study, titled "It Was 80% Me, 20% AI", included 19 fiction writers, 30 readers, and AI-generated suggestions using OpenAI's GPT-4. Researchers sought to look into the apprehensions regarding how generative AI could change the writer economy, especially about the authenticity of creative work.
Lead researcher Angel Hsing-Chi Hwang explained that for an author or writer, the value of someone's work is what it means to be authentic. In this regard, co-writing with AI might destroy this purpose.
Read More: Top 5 AI Breakthroughs of 2024
Researchers Used CoAuthor in the Experiment
Participants used a program called CoAuthor, developed at Stanford University, to write 200-word passages. CoAuthor integrates with a large language model (LLM), allowing writers to generate and insert AI suggestions by simply tapping their keyboard.
Writers completed two versions of the task.
- Using GPT-4 without any personalization.
- Using a personalized version of GPT-4 trained on samples of their prior work, employing in-context learning techniques.
Interestingly, the writers were not briefed on which copy was personalized in order to not influence their comment on the process.
Authenticity and AI: Writer Issues Emerge
According to ZDNet, the experiment showed a little bit of resistance and fear toward using AI, with writers asking questions such as:
- Loss of originality: The writers expressed that AI could take them out of their box and lead them to produce much lower-quality work.
- Reduced pleasure and mastery: Some complained that AI might take away the emotional satisfaction of writing.
- Ownership issues: The writers were concerned that work generated with the help of AI could not be authentic and truly representative of their personality.
One of the participants summed up these concerns by saying, "Authentic writing is about the story told from the heart of an author. Can they get inspiration from AI? Absolutely. But it's not the same."
Personalization and Writer Preferences
Despite initial skepticism, many writers preferred the passages generated with the personalized GPT-4 model once the differences were revealed. Personalized AI tools, which incorporated the writer's tone and style, appeared to strike a balance between inspiration and authenticity.
Researchers noted that personalization helped writers feel a stronger connection to their work, alleviating some concerns about losing their creative voice.
What Does 'Authenticity' Mean for Writers?
It identifies two key aspects of authenticity.
- Authentic self: This pertains to a writer's ability to make a case for their identity and ownership of a piece. One participant described this as, "Authenticity is about producing something that only I could create."
- Authentic process: Here, writers felt emotionally invested and were fully in control during the creative journey.
AI in Creative Writing
While generative AI tools such as GPT-4 undoubtedly have their benefits—speeding up workflows, sparking new ideas, and more—they also pose fundamental questions about the nature of creativity. Doubts on the part of writers suggest the need for AI tools that enhance, rather than overshadow, the unique voices of writers.
AI can become a collaborative partner instead of a disruptive force by respecting individual style and personalization.
Now, we might finally know the reason why global screenwriters' guilds are against AI training data.
