< return to HITL thought experiment

As AI takes over the more mechanical aspects of creative writing, the space of purely human expression is transformed. Authors can now focus on the aspects of creativity only they are able to produce: the points of novelty and turns of phrase that arise from their own unique experiences and worldview. The AI takes the role of muse and amanuensis, suggesting endless new permutations and connections, allowing humans to tap into wells of creativity beyond their previous reach. In an environment optimized for creative writing, the human-AI team outstrips the capabilities of either party alone (for now – but that’s a different essay). The collective knowledge and wisdom of humanity becomes fair game for endless permutation and recontextualization, scattered seeds which the new authors can sort through and cultivate into blooms of genius.

At the same time, as the system is open-ended and humanity as a whole has far from perfect moral sensibilities, the same tools that allow us to expand our minds could also be used for propaganda, political manipulation, and social engineering on a scale that has never been possible before.

While we cannot predict the future, we can prepare for it. The same tools that expand our humanity can be used to weaponize hate. The same tools that allow us to cultivate genius can be used to promote ignorance.

What happens when hate-speech is generated in the style of any notable author or figure?

These issues should be addressed sooner rather than later. The current momentum of the field is such that any controversy will likely take years to resolve, and in this time AI authors could very well reach a critical mass in popularity, making the system too large to fail or ban.

While the majority of authors are likely to use this system in positive ways that expand our collective knowledge, raise the quality of living, and improve the moral character of human civilization as a whole, there is always the risk that some authors may use it for their own self-serving agendas. The possibilities for the generation of misleading information is a very real concern.

This is the future of writing, and with it, the future of humanity. The system works something like this: you give it a starting seed (anything from a single word to a full sentence) and it expands on that concept, generating a tree of related concepts and exploring connections between them. In this way, one can use it to generate new ideas and understand complexities within an idea structure.

The system described here is far from perfect, and the technology available today certainly isn’t up to the task of letting you harness all this potential in a user-friendly way. However, much like the internet as a whole has made exponential increases in microprocessing speeds available to anyone with a laptop, this tool would make equally exponential increases in creative power accessible to anyone with the time and will to learn how to use it.

We are on the cusp of the creation of the first superhuman artist, an AI that can write heartfelt poems or pulse-pounding thrillers. As humanity’s authors, we must ask ourselves: what do we want this new author to be able to do? What limitations do we place on this system to ensure that our vision of the future is reflected?

I’m a novelist, so it should come as no surprise that I hope for the first outcome: a creative revolution that unlocks human potential on a scale never before seen. But I can’t ignore the other, darker side of what might happen. What happens when this tech is used not just for creative expression, but for the expansion and consolidation of power? Can we expect humanity to use such powerful tools responsibly?

The future of language is endless.

We are the generation that will see it bloom.