Keep your witchy wits about you while researching online!

BryWisteria · July 15, 2026, 12:20pm

Merry meet!

Just wanted to slide a reminder from someone who does a lot of research searching and digging through the internet for my fellow witches doing research or fact-checking. Many search engines these days have an auto search result at the top. It is often not correct.

Here’s what ruffled my feathers: As many of us who study astrology know, while the typical start/end dates for astrological seasons are set, the exact day the Sun enters a sign (and starts a season) varies from year to year. I always like to double check across astrology sites to make sure the current dates for the year are correct in my notes.

I ran a quick search for “sun enters Leo 2026” in my search engine (DuckDuckGo) and got the following at the top of the page:

The “search assist” claims that the Sun enters Leo on August 17. This is not true, and right below the search assist you can see the actual results displaying the correct information. The standard Leo season dates are estimated from July 23 to August 22 but this year, the Sun enters Leo a little before that on July 22.

Ironically, the search assist is claiming that it used CafeAstrology as a source - so not only is the search assist wrong, it is also displaying false credit info

I took a look at the other source listed, and sure enough, there are the false dates. Now, maybe this site is using a completely different form of astrology from what the vast majority of folks searching for Leo season would be looking for (and thus what the first search result should display), but I couldn’t find reference to that on this page. The article also shows signs of being written by GenAI. I’m thinking it was likely machine generated and not fact-checked.

This is not the first time I have seen search assist be wrong, but I often see people taking and using what it presents as fact.

All this to say: When researching online, please be mindful where you put your trust! Search assist is not reliable and needs to be fact-checked. Some websites cranking out GenAI content are not fact-checking their work and the resulting misinformation can range from wrong to harmful. This is becoming more and more of an issue as GenAI articles and content flood the web. AI is feeding off of AI and making very muddled content.

Please know I’m not saying this to be scary or (heavens forbid!) deter people from studying what they love. I’d just like to renew everyone’s awareness and encourage mindful research. Stick with content and sources you trust whenever possible.

Keep on exploring, witches. Happy studies and blessed be!

Moonbeam · July 15, 2026, 2:48pm

@BryWisteria

Thank you for sharing such important information. When I put my icon and the phases thereof I always check the Spells8 Homepage on the Daily Lunar Spell and Merry Meet Monday for other events as even though the time zone might be slightly different by a few hours I am posting as I am in the US. As it would be near impossible to get the precise time and I do not want to have to keep looking it up and it can be confusing for everyone here.

My search engine is IPhone Safari which takes me to Google which as you say automatically fetches up Top Search Results given that these are generated by AI and I like to use the old ways Forklore which is what I have learned and passed to me through the Hedge AI is not really relevant thank goodness for that!! . However, I know that this is going to be the new normal for almost everything in daily life so I am mindful of what information is being generated I do like to use my own brain but there will be times that I will be guided by it as some information will not be accessible by any other way. I have chosen to keep it out of my craft but do understand that research and assistance by AI is done on the Forum but have also noted that it has been highlighted in the relevant post so I am aware of it.

Much love and many blessings

Feathertip · July 15, 2026, 3:14pm

Thank you so much for this reminder, Bry! It’s important to remember how AI has inserted itself into all of our search engines.

I don’t think it only inserts itself into DDG or Safari or Google - it’s become omnipotent, and it is unfortunately our new reality when going online. Even a casual query will see those obnoxious “Search Assist” words pop up as the very first thing we encounter - which might lead some of us into believing it’s accurate!

It’s something I’ve noticed recently, and now I make a real effort to pass by that so-called “top-tier” reply, and just scroll past it and down to go to other sources. As you point out, some will be correct, and others might still not be - it’s on us to look carefully and keep out “witchy wits” about us!

AI is everywhere, and is really not to be trusted as a reliable source - it isn’t. Thanks again.

georgia · July 15, 2026, 4:26pm

A friend once told me, before AI took over, that when she researches, she verifies it with three reliable sources. Reliable sources are tricky because even they can be hacked or have older information which has been learned more about in intervening years.

The movie Thirst quoted the Mayo Clinic in saying that Thrush is the white “jacket” on the tongue which is caused by lowered intake of water and will heal when body hydration is returned to normal. Our lovely @MeganB pointed out the error in that. Thank you, dear. Simply a case in point that it can be dangerous to take anything written as “Gospel Truth.”

Moonbeam · July 15, 2026, 4:42pm

@BryWisteria @Feathertip @georgia - it just seems more work having to verify everything time consuming and exhausting I will have to I expect!

You are spot on with that I know that my Employer is already using Copilot as over last month this button has been shown on my computer screen and this month there have been training on AI. It was just for Solicitors not Para Legal or Legal Secretaries but I know they will extend the same in time. Everything as you say is online to making any appointments definitely banking, local authorities, utilities. My hairdresser prefers appointments made in person and as my best friend does my beauty as she owns her own Salon I am fine with that. There is an option for online grocery shopping but I will always go to the Grocery Store to pick my own as they are in business to sell the oldest products and produce first and I look at the dates. i cannot see that grocery would stop this option as they are of the view that you will purchased more in store. I also don’t use delivery APPS which took off during Covid as again I prefer not to.

I do know that Pandora’s Box is well and truly open and it cannot be put back and reluctantly in some areas the choice has been taken away.

On the brighter note as I always try to be positive we can still grow somethings, herbs, vegetables even in a large pot and help Mother Earth and nature the best we can. Just harvested my second early potatoes.

Blessed Gardens

leilani1 · July 15, 2026, 5:37pm

Witchy Wednesday Bry,

Thank-you for bringing this up. So many believe that if they read it online it has to be correct. I remember, back in the 60’s we were told that if it was on the news it was a fact, just like if the doctor gave you a pill, it would work. Too many people put blind faith in too many things. “Always do your homework”

Lady Gene

Moonbeam · July 15, 2026, 7:53pm

@leilani1 very true Lady Gene

Amethyst · July 15, 2026, 10:09pm

Thanks for this! Seems like AI isn’t as good as everyone is saying!

Feathertip · July 15, 2026, 10:49pm

You know me; I ain’t saying it.

dan3 · July 15, 2026, 11:22pm

As a frequent AI user, I tend to think the AIs are repeating back things they discover - not purposely making stuff up. There are exceptions to that (Bry’s example - which seems to be a real mess up).

Lately, most AIs say they can make mistakes - which they do. They tend to reach out to websites and pick the information that matches requests. That information may be flawed.

I recently did a search on whether a prominent news woman was pregnant (I read it on a gossip page). The AI came back that she had 9 kids and expecting a 10th. That was so obviously wrong, I searched the web normally and found an obscure East Indian site that said it. I think the AI couldn’t find anything and took the only information that matched my query.

This weirdness actually results in some AIs having conflicting information and answers about identical questions. For example, ask an AI if any of Hel’s undead participated in Ragnarok. The answer is most correctly stated that there is disagreement about it - AIs tend to give either a ‘yes’ or ‘no’ answer.

I’ve had AIs swear they knew about a book series (I ask them before I ask my question) and then not know about key characters.

So… I guess I said that to say - any favorite AI has weaknesses and just like searching the old - non AI - web, verify, verify, verify.

Oh… and don’t ask me about statistics calculations. They mess up math a lot. I stopped asking them.

Finally - AIs are part of the Internet. As a famous person once said:

leilani1 · July 16, 2026, 1:04am

You have me on the floor. JFK and the internet?!! Lady Gene

BryWisteria · July 16, 2026, 11:14am

I think you’re right that it is the new norm, and I agree - the best way to proceed is to do just that! A little mindfulness and some extra awareness about the content we consume goes a long way

I do the same! To be honest, it’s a little hard - the box is so bright and shiny and they make the words appear with a little animation. It’s very eye-catching and I can understand why some folks might just stop there! But you’re absolutely right. I think scrolling right on by it to check the true search results ourselves is a better and more accurate way to get our information

You raise an excellent point, Georgia! Misinformation has always been out there, and perhaps a timeless bane of academics. I can only imagine how hard it was to study new things back before the internet or even before the printing press, when books were one-of-a-kind and extremely rare resources. A student would be very limited in the resources they had at hand and even more limited in their ability to cross check them against other sources. It makes my head spin a bit to think about it!

The modern day predicament with cross-checking sources isn’t nearly as challenging as it was in ancient times, but I do think AI will continue to accelerate the amount of misinformation out there. And perhaps even more challenging is the speed and convenience in which AI presents false information to knowledge seekers.

Sorry, I’m rambling now. I completely agree with you about being mindful of misinformation in any form. I like your friend’s method of triple checking with reliable resources!

Love this! No matter how omnipresent AI becomes, you’re right - at the end of the day we still have the power of choice. The option to set it aside however and whenever we can is within our power. And choosing nature is always the better option (at least in my eyes!)

Your potatoes looks amazing! What a great harvest

I was never very close to my grandparents, but I’ve heard that all four of them held this mindset to a fault. My parents, aunts, and uncles often make reference to how my grandparents would do whatever the doctor told them, even if it was strange or perhaps not an appropriate medical diagnosis. Blind faith in anything is truly a dangerous thing!

Well said!

My pleasure! I know I appreciate reminders about things from time to time, so I thought I’d try to pay it forward

This is true! Beneath the DuckDuckGo search assist (and I assume other AI searchers as well) there is a little note that says “auto-generated based on listed sources. May contain inaccuracies”. So it’s good that they are upfront about it

That makes sense, and is a good thing to be aware of when searching for rare or hard-to-find questions. It seems like the more challenging the question, the more critical one should be of any auto-generated results, knowing that its resources are limited.

This is why I was so surprised that the dates for an astrological season were presented incorrectly. It seems like a rather common and generally well-known/well-searched for topic! Although that may just be me being biased about my own interests

Three cheers to this! I agree - it may take some extra time and effort, but the more reliable sources that verify the information the better!

Really!? I’m surprised to hear this! I often say that, while I believe creative works should be done by human creatives, perhaps there is a place for AI with handling calculations. It seems like something any computer would be very good at doing (and probably much faster than humans can calculate on their own). I wonder why it struggles? Perhaps it has something to do with how AI models draw on other resources, instead of doing their own calculations? I know you’re very smart with tech stuff, Dan, so if you hear anything about this (or have your own theories!), I’d love to hear them!

dan3 · July 16, 2026, 3:49pm

I have my own guess!

Statistics is very particular about how you pose a question and there are SSOOOooo many ways to pose a question depending on what you are wanting.

I think the AI occasionally gets confused with interpretation… and ends up making obvious mistakes.

Simple calculations (measures of central tendency, etc) are usually fine (always fine in my experience). More difficult calcs (probabilities using subsets of populations with order important or not important) have to be carefully considered.

Having said that, I just went back to a statistic question I had several months ago that an AI got wrong. Was going to use it as an example. Today, The same AI is getting it right. It could be that AI growth has overcome that particular issue.

Amethyst · July 16, 2026, 10:01pm

Is there any way someone could do, like, a screenshot of something when it’s AI, at least on Amazon? It’s difficult for me to tell sometimes. Where do you look? I can spot the obvious videos; it’s the not-so-obvious ones that get me.

Although I’ve got to admit, AI has made it much easier for crypto videos. Mothman has never looked so good!

BryWisteria · July 17, 2026, 12:51pm

I see! That aligns with what (admittedly little) I know about how AI works - it seems like it is good at following clear instructions step-by-step to complete a larger task. But since AI can’t actually think for itself (and is limited to the data it can read/draw from), it would have to “guess” about making interpretations or doing anything that requires a bit of abstract thought. While humans can “read between the lines”, AI can only take the “lines” (data) and work with them at face value.

It’s interesting to think about and try to understand. Thank you for sharing your thoughts and wisdom about it, Dan! I know you know a lot more about it than I do, so it’s appreciated

It’s very possible! The models are continuing to be developed. I used to look for the six finger trick as it used to be a clear giveaway that an image was AI, but it seems like most models have advanced to where they can usually manage five fingers now - at least on realistic/photograph style pictures.

You’re not alone! It is becoming harder and harder to spot AI. At this point, rather than look for one clear giveaway in text or a photo, I think it’s best to look for multiple “odd” or “strange” things that, when put together, could give AI writing or images away. I follow an artist on YouTube and she made a video about methods someone can use to check to see if an image is AI. Here is her video:

For writing, it’s unfortunately a lot harder - especially in shorter pieces. There was a famous test run by a group of authors where they posted 10 pieces of flash fiction (under 350 words) and readers voted to see if they could tell which pieces were written by humans and which were generated by AI.

(The Can You Tell if Writing is AI or Human Test is here - you can still take it if you’re interested!)

The results were… not encouraging. And the original test ran in 2023, so AI models have gotten a lot better at hiding their robotic give aways. That being said, things become more obvious in longer pieces of writing when AI begins to forget about previous plot points, character names, etc. Although at that point, a reader has already invested a lot of time and effort into the book so …

For Amazon, I’d look first at the book cover. If the cover looks like it was made with AI, it means the author is using AI - so it’s very likely they also used it for the writing. You can also check the reviews and see what other folks think.

Whatever the medium, I think it ultimately comes back to being as mindful as we can be, and double checking other resources/reviews/etc to help us make an educated decision about what content we put our trust in

leilani1 · July 17, 2026, 1:57pm

Bry, as my degree is in History, I am used to fact-checking. Honestly, though, history is written by the victors; so everything we learn as fact, needs to be researched. Just look at this, really simple…back in the 40’s cigarettes were touted as “relaxing” so they were good for you…NOT, cancer. So, just take everything with a grain of salt. (just that little bite in there, to get you thinking)

Lady Gene

Amethyst · July 17, 2026, 10:46pm

Thank you for all that good information and the video! I’ll be going through it and looking more closely at things online!

dan3 · July 19, 2026, 3:58pm

Early on, I spent lots of time fixing images - at times enough work was put in for me to want to be a co-artist. It took weeks of work on some decks. I’m still finding them when I use one of the older decks.

So very important to realize. Even the basic search engines are automatically using it - and the links provided may also use it and not say anything.

@Moonbeam said it is the "new normal’. We can carry that to the bank. It is not a case of being discriminating at the front end - clean results when searching are up to us… not them.

I still love AI!

Feathertip · July 19, 2026, 11:43pm

Well, you do a lot of work with it - that’s a positive thing, as opposed to having it step in front of every search engine out there. The craft aspect of it isn’t any worse than photo software.

Moonbeam · July 20, 2026, 5:52am

@dan3 yes we can have the choice to scroll past