"I have affirmed the premise that the enemy can be so simple as a bundle of hate," said he. "What else? I have extinguished the light of a story utterly.

LaMDA was not designed as a writing tool. LaMDA was explicitly trained to respond safely and sensibly to whomever it’s engaging with.

Only through open and ongoing dialog between technologists and artists can we build tools that have a positive impact on the world.

The novel workflows that a technology enables are fundamental to how the technology is used, but these workflows need to be discovered and refined before the underlying technology can be truly useful.

Can a language model be transgressive without intentionality?

More provocatively, great writing is transgressive — it subverts expectations and challenges the reader.

LaMDA's safety features could also be limiting: Michelle Taransky found that "the software seemed very reluctant to generate people doing mean things". Models that generate toxic content are highly undesirable, but a literary world where no character is ever mean is unlikely to be interesting.

A recurring theme in the authors’ feedback was that Wordcraft could not stick to a single narrative arc or writing direction.

A recurring theme in the authors’ feedback was that Wordcraft could not stick to a single narrative arc or writing direction.

Several participants noted the occasionally surreal quality of Wordcraft's suggestions.

One of the most well-documented shortcomings of large language models is that they can hallucinate. Because these models have no direct knowledge of the physical world, they're prone to conjuring up facts out of thin air. They often completely invent details about a subject, even when provided a great deal of context.

Writers struggled with the fickle nature of the system. They often spent a great deal of time wading through Wordcraft's suggestions before finding anything interesting enough to be useful. Even when writers struck gold, it proved challenging to consistently reproduce the behavior. Not surprisingly, writers who had spent time studying the technical underpinnings of large language models or who had worked with them before were better able to get the tool to do what they wanted.

Many authors noted that generations tended to fall into clichés, especially when the system was confronted with scenarios less likely to be found in the model's training data. For example, Nelly Garcia noted the difficulty in writing about a lesbian romance — the model kept suggesting that she insert a male character or that she have the female protagonists talk about friendship. Yudhanjaya Wijeratne attempted to deviate from standard fantasy tropes (e.g. heroes as cartographers and builders, not warriors), but Wordcraft insisted on pushing the story toward the well-worn trope of a warrior hero fighting back enemy invaders.

novel is not the same as interesting

Wordcraft tended to produce only average writing.

“[It's] an amazing tool for brainstorming or rubber-ducking. Its conversational quality is perfect to talk about plot, characters and worldbuilding.”- Nelly Garcia

We've also integrated a chatbot feature into the app to enable unstructured conversation about the story being written.

“...it can be very useful for coming up with ideas out of thin air, essentially. All you need is a little bit of seed text, maybe some notes on a story you've been thinking about or random bits of inspiration and you can hit a button that gives you nearly infinite story ideas.”- Eugenia Triantafyllou

The authors agreed that the ability to conjure ideas "out of thin air" was one of the most compelling parts of co-writing with an AI model.

language models are incredible "yes, and" machines, allowing writers to quickly explore seemingly unlimited variations on their ideas.

Wordcraft shined the most as a brainstorming partner and source of inspiration. Writers found it particularly useful for coming up with novel ideas and elaborating on them. AI-powered creative tools seem particularly well suited to sparking creativity and addressing the dreaded writer's block.

The workshop cohort consisted of 13 professional writers, who were given 8 weeks to use Wordcraft. Unlike more constrained user studies run in the past, we gave writers freedom to write anything they wanted using any workflow they could devise. We share below some of the insights we learned through this process.

In addition to specific operations such as rewriting, there are also controls for elaboration and continutation. The user can even ask Wordcraft to perform arbitrary tasks, such as "describe the gold earring" or "tell me why the dog was trying to climb the tree", a control we call freeform prompting. And, because sometimes knowing what to ask is the hardest part, the user can ask Wordcraft to generate these freeform prompts and then use them to generate text. We've also integrated a chatbot feature into the app to enable unstructured conversation about the story being written. This way, Wordcraft becomes both an editor and creative partner for the writer, opening up new and exciting creative workflows.

In addition to specific operations such as rewriting, there are also controls for elaboration and continutation. The user can even ask Wordcraft to perform arbitrary tasks, such as "describe the gold earring" or "tell me why the dog was trying to climb the tree", a control we call freeform prompting. And, because sometimes knowing what to ask is the hardest part, the user can ask Wordcraft to generate these freeform prompts and then use them to generate text. We've also integrated a chatbot feature into the app to enable unstructured conversation about the story being written. This way, Wordcraft becomes both an editor and creative partner for the writer, opening up new and exciting creative workflows.

For instance, if the user selects a phrase, a button to "Rewrite this phrase" is revealed along with a text input in which the user can describe how they would like the phrase to be rewritten. The user might type "to be funnier" or "to be more melancholy", and the Wordcraft application uses LaMDA and in-context learning to perform the task.

We like to describe Wordcraft as a "magic text editor". It's a familiar web-based word processor, but under the hood it has a number of LaMDA-powered writing features that reveal themselves depending on the user's activity.

The application is powered by LaMDA, one of the latest generation of large language models. At its core, LaMDA is a simple machine — it's trained to predict the most likely next word given a textual prompt. But because the model is so large and has been trained on a massive amount of text, it's able to learn higher-level concepts.

Our team at Google Research built Wordcraft, an AI-powered text editor centered on story writing, to see how far we could push the limits of this technology.

First, I’m impressed as hell by the Wordcraft team. Daphne Ippolito, Ann Yuan, Andy Coenen, Sehmon Burnam, and their colleagues engi­neered an impres­sive, provoca­tive writing tool, but/and, more importantly, they inves­ti­gated its use with sensi­tivity and courage.

Prioritizing thinking or wordcraft is an intriguing way to divide writers.

Jay Clayton