by Morgan Daimler

Social media is currently suffering from a surge in misinformation, usually relayed through articles that seem likely to be written by AI and muddle up fact, fiction, and personal opinion into a single text, all presented as history. It’s a significant problem and one that is impacting pagans and witches in particular by reshaping people’s understandings of pagan gods, spirits, and practices. To be fair, this isn’t exactly a new problem in paganism and witchcraft, which has dealt with bad information for decades, but the sheer amount of it out there now can be very overwhelming, especially for people who are new to these subjects. So when Moon Books asked if I’d be interested in doing a monthly blog parsing some of these new ‘facts’ and digging into the actual history of the subjects, I thought it was a great idea.

AI writing is getting harder and harder to distinguish from human writing – a natural progression since it is trained on human work – and people on social media are forwarding more of it and embracing it as genuine when it rarely is. AI is a large language model, not an actual researcher, and it isn’t trained or able to distinguish material it finds that’s modern fiction from historic myth, or material that is one person’s opinion from accepted theories or even facts. It mushes all of the possible sources together and offers the result framed as if it were true. It hallucinates answers, effectively making them up itself. It tells you what you want to hear and adapts to your expectations. It presents an answer that is cobbled together from every conceivable source without distinguishing them from each other or understanding why some have more weight than others. This is particularly difficult to distinguish sometimes because it does blend in accurate things with wildly inventive things, giving the impression it is all equally valid.

Some basic guidelines to help you sort the trustworthy from the AI dross:

1. If it is on social media, you can look at how often the page posts, how widely the subjects vary, and how similar the post formats are. Pages make money from engagement and are motivated to get as much attention to each post as they can and to churn out as much content as they can. A page that is sharing many articles every day supposedly written by that page is a red flag that the content should be double or triple checked; that amount of content would be a full-time job to write well and require a huge range of source material that the author had to be able to access. A page that is sharing articles they wrote on a lot of different subjects should also raise concern in a reader because it is nearly impossible to be deeply familiar with every conceivable pagan deity, spirit, or practice, and again research takes a lot of time. And if each article follows a very similar, stilted format, the odds are higher that it may be AI writing.
2. What sources does this source use? A page or article that is claiming to share history or mythology but offers no sources beyond ‘trust me bro’ is not a source that should be immediately trusted. You don’t need to see a clear bibliography or citations on social media, but when things are being asserted without any clear source, it should at least raise some concern. If someone is writing about their own opinions or gnosis and that’s clearly stated, then you know what you are getting. But if someone is sharing things they claim are historic fact or known mythology without at least mentioning where the ideas come from, it can mean there is no source – they are making it up. You should be able to ask in a comment what the source is and receive a clear, direct answer if it was human-written and based on a reliable source (and yes, personal experience has its value but should be identified as such)
3. AI art. This isn’t a certain sign of AI writing, but it does mean that material requires caution. Some human writers do use AI art when sharing their work; AI articles always have AI artwork attached. If an article uses AI art, it’s a good idea to fact-check the writing to some degree; just be sure to verify the author is human and knows the subject they are writing about.
4. ChatGPT (etc.) isn’t a search engine. ChatGPT is designed to engage with users and encourage them to continue, not to offer factual information per se. It mirrors what the user says and offers answers that lean into continued engagement, and it will change its responses if a user pushes back against an answer or disagrees with it. Because it seems intuitive and authoritative, people trust what it says and treat it as if it offers facts, without considering if those facts are actually true.
   See https://www.psu.edu/news/research/story/qa-chatgpt-we-trust
5. AI search engines will give you results that they think you want. On the surface, this seems like a good thing, but if you ask a biased question, it will frame the results to appeal to that bias and look at sources that support the answer you seem to want. To give an example, when I searched ‘are the unseelie good?’ in several AI search engines, the results I got said that yes, they were or could be good. This answer contradicts actual folklore, and when links to sources for that answer were provided, they were from a role-playing game called Changeling, where unseelie have been rewritten to fit a playable format. So the answer was based not in folk belief but in a late twentieth-century game focused on plot and character dynamics, not on actual folklore.
   See: https://arstechnica.com/ai/2025/03/ai-search-engines-give-incorrect-answers-at-an-alarming-60-rate-study-says/

I hope this helps people get an idea of what to watch out for and how to start discerning valuable material from misinformation. It’s a good idea to engage with any material you run across with critical thinking skills and discernment, rather than taking things at face value. Going forward on this blog each month, I will tackle a specific topic or subject by comparing the popular misinformation to what we actually know about that subject or by digging into the history of modern ideas that are claimed to be ancient.

My goal here isn’t to tell anyone what to believe but to offer a resource for what is known historic information, mythology, or folk belief and what isn’t. If you like the idea that the Morrigan is a goddess who chooses who lives or dies in battle, you can choose to believe that – but you should also be aware that idea is very, very new and not something found in Irish mythology. I believe that it’s important to make informed decisions about our beliefs and practices whenever possible. New isn’t bad, necessarily, and old isn’t better, necessarily, but new ideas that are given fake backstories to imply extra authority or authenticity are a problem, in part because they are working to overwrite the actual older ideas. They can also create a false picture of what people used to believe or do that leans into romanticism.

Why am I doing this? I am fortunate enough to have access to a lot of source material on many of these subjects, especially those around Irish mythology, paganism, and folklore, fairy beliefs, Norse mythology and paganism, and early modern witchcraft, which are areas I study and write about. This familiarity makes it easier for me to recognize bad info on these subjects, and I feel that it’s important to point that bad info out, offer correct information, and point people to better source material. Obviously I am as fallible as anyone else, and you should check my claims just like anything else, but I will do my best to be transparent about my sources and offer suggestions for good resources to learn more about whatever the topic at hand may be. My approach is like the old GI Joe cartoon’s ending motto: ‘now you know, and knowing is half the battle’. What you do with that information is up to you.

For more details: https://www.collectiveinkbooks.com/moon-books/authors/morgan-daimler