« on: April 13, 2024, 03:54:01 PM »
Large Data Models & Generative "AI": Our New Policy
As many of you know, machine learning based tools for generating content from large data models have been significantly expanding in their usage in recent years. ChatGPT, Midjourney, Stable Diffusion, and other text and image generating services are increasingly prominent in popular discourses about art, games, and the future of creative works.
The voting members of Exilian have recently voted on a new policy for how we handle this sort of content. We recognise that it's something people are going to want to discuss and play with, and we want to assure people that discussing these technologies isn't unwelcome here. However, there are a lot of concerns about generative AI, including but not limited to their potential for disinformation, the effect on artists and writers whose content is taken and reconstituted by these systems, the environmental impact of very heavy data model usage, and the issues of huge quantities of automatically generated material flooding the internet. The legal situation around generative AI is also still unclear and it's important that we protect this community from any potential legal issues. We don't think that we can simply ignore those problems, and we want to keep Exilian as a space that supports and cherishes human creativity first and foremost.
In future, that means that Exilian will not be providing formal support to projects that use generative AI. That includes not giving newsletter space or social media shoutouts to such projects, and it also means that generative AI pieces will be explicitly banned from entry into future competitions and site events. We also have some new expectations about posting generated content: that people should keep it limited and not flood the site with such content, that you mustn't post work that explicitly mimics specific writers' or artists' styles, and that if you're posting content produced by an AI then you have a responsiblity to clearly label it as such.
What we can do on these issues as a small community is limited: "AI detectors" are themselves largely machine-learning systems and have significant flaws, and whilst we will take what steps we can to avoid content you post on Exilian getting scraped by large generative model providers, as a public-facing forum without a legal budget there are limits to how much we can prevent machines reading content on here. Nonetheless, we're doing everything we can on these issues to make sure that we keep Exilian as a space that will support our artists, game developers and writers in their own creative endeavours, and we hope that taking these steps will help with that.
You can read the full text of our Generative AI policy here.
With best wishes,
The Exilian Team
Logged
The duke, the wanderer, the philosopher, the mariner, the warrior, the strategist, the storyteller, the wizard, the wayfarer...
