Note: This is not a legal document. It is an open letter to our users for the sake of transparency on how we (presently) interpret what is and is not an ethical usage of AI, and our commitment to the former.
AI has taken the world by storm and caused a great deal of anxiety for creatives. It certainly doesn’t help many authors have had platforms suddenly inform them their data will be used to train models without their consent, AI books and art are now filling up Google searches, writers have lost their jobs to ChatGPT, the energy demand of AI is now a climate change contributor… And yet, there are also those who acknowledge AI is extremely useful in assisting with rotework–all those things that get in the way of doing things you actually enjoy doing. Furthermore, it can help with complex data analysis that humans are simply not capable of on their own terms.
Basics of AI Ethics
There are many AI ethical principles/frameworks. Which one a company finds most suiting for their situtation will vary — assuming they have put any thought or effort into choosing one to mitigate the consequence should AI cause harm to both real people and their own company/organization.
Many of these share the following principles:
- Privacy
- Accountability
- Transparency & Explainability
- Fairness & Non-discrimination
- Safety & Security
- Human Agency
- Do No Harm / Respect for Human Rights
There are also standards and regulations that vary from country to country (and state by state) on how data is stored, handled, used, and so forth.
What we do
- As a base rule: we place a high value on integrity and can morally be described as negative utiliarian stoicism. This means we’re more concerned with non-maleficence than making everyone happy. What choices we make will follow accordingly.
- We sometimes use AI to help with rote tasks like list compilation. As much as we would love to know the name of every single indigenous group to ever exist, or otherwise have days of our life to manually to research and record a monolith not even Wikipedia contributors have managed to create — we are but humans who would rather write in third-person plural about how our failure to be omniscient gods gives us anxiety.
- We sometimes use AI to help with editing, writing tests to catch bugs, fixing bugs (where answers not easily available online), and rote tasks like refactoring (which then usually have bugs that need a human to manually fix). This helps to drastically reduce our time spent on figuring out how to fix stuff and more time on actually improving our application. That said, despite our best efforts to make our app bug-free as possible, it does not mean a bug-free platform is possible because there will always be bugs and that also gives us anxiety.
- We use images from the public domain in our website media. While we try to avoid it and do not generate AI images ourselves, some of the images we chose may be AI generated until we have better resources to replace them.
What we DO NOT do
- AI is not presently integrated into our platform. If this changes in the future, we will continue to maintain transparency and keep open dialog with our users about their ethical concerns.
- We DO NOT COLLECT DATA about your usage. This unfortunately means we do not/have no plans to create a “recommendation algorithm personally catered to you.” When you enter your filters, the books you see are genuinely random. This means you may see books you don’t like or would have never considered the possibility of existing. That’s kind of the point. But in this case, we recommend adding more filters or putting the phone down to actually read some of the books you’ve added to your stack. We believe this approach improves transparency, reduces automation bias, and maximizes human autonomy.
- We DO NOT ALLOW AI WORKS on our platform. Creatives may only add works you have personal copyright claim, and that means works you have written yourself. For more on this, please refer to our Terms of Service.