AI - The Genie is Loose

Second Post of the Day

Most people know that I do NOT like AI. In fact, I detest AI and do not use it. 

Can I be fooled and miss something that is AI? Of course - AI can be quite sophisticated.  But I can usually recognize its use when I am doing research or writing, be it for a blog entry, sewing project, or something else.

So, with such a distaste for AI, can you imagine my dismay when I got this email?

My reaction was instantaneous and I said aloud, "Oh, no you don't!"

Academia.edu is a reputable site for educational researchers. We use it for publishing and storing papers that we have written or for finding papers to use as we do research for papers that we are writing. It's a respected site, though not quite not as big or essential to the educational community as Research Gate.

Regardless of how reputable the site is, using AI on one of my papers was a serious issue to me. I opened the email to see more about what Academia.edu was doing. My dismay grew when I read the email.

An AI had created a podcast with one of my papers? No, no, no, no, NO! 

As I looked closer and saw the length of the podcast, I could only wonder what was included in that podcast. The podcast was just 4:45 long; that paper could not be read completely in that amount of time.

I was curious but I was absolutely not going to click on the link to hear it. And, while I am glad that this so-called podcast is private to me, I am NOT happy that it has been created.

This particular paper generated a lot of interest when it was first published in Research in Science Education in 2002 and it has been heavily used by other researchers for years. My guess is that Academia.edu selected it for that reason, perhaps hoping to make it more usable by researchers who want the information and prefer the podcast approach.

Whatever their reasons, though, I am not happy about discovering AI being used with my work. AI is creeping into everything, everywhere. 

We truly have let the AI genie, for good or for bad, out of the bottle.

A Shock from Artificial Intelligence

I don't like AI and the assumptions it makes. No, I do NOT want push notifications for a gajillion sites. Yes, I know how to set those. No, I do NOT want reminders that I haven't answered an email. No, I don't want to nudge the person who hasn't responded to my email yet.

And so on.

But AI took the cake the other day. And I was absolutely shocked.

I am the one who found Abby and Two Socks dead in the pasture last June. For a variety of reasons, possible insurance reasons being one of them, I took a photo of Abby as she lay in the grass. 

It's MY photo and I have never shared it with anyone or even told anyone I took it. It's highly, highly personal and no one else's business. When it was clear that I would not need proof of her death, I could not bear to delete it. That was my sweet girl lying in the grass. My last moment with her.

I never look at it although it is in my photos on my phone. If I happen to scroll past it, I don't stop. That was a terrible day, and while I cling to her last picture, I don't want to look at it. Its presence is all I need.

Well, several pics popped up on my phone yesterday - suggestions from AI for a new phone wallpaper photo.

The first suggested picture was the one I'd taken of Abby.

I was shocked, I was livid, I felt sick, I cried. My privacy and my grief had been violated.

I am completely happy with my current wallpaper. Yes, I know it has been there for years. No, I don't care about varying it. No, it is NOT out of date. It's a beautiful picture of my girl and a favorite of mine.

I make that choice, AI, and I don't need your suggestions. But here's a suggestion for YOU:

Add some sensitivity to that ridiculous artificial intelligence you supposedly have.

Determining AI - Please Educate Us

Some time ago, a photo was posted on social media that upset more than a few people. Apparently it had been generated using AI and the discussion quickly became an angry one with some folks calling for the person who had posted the photo to be banned from the group.

The discussion then broadened to ban anyone who used/might use AI, posted pictures that were AI generated, and so on.

Gulp.

"She/he should have known" is not always true. That original poster easily could have been me. (Although it was not. Nor was it anyone I know.)

Personally, I could not tell that it was an AI picture. Just as I sometimes struggle with determining if a model is made from chalky plastic (I can tell a base coat without problem), I struggled with trying to see where the indicators were of that photo having been generated by AI. I have read advice and comments before about how to do so but I simply cannot see it. 

The reasons those who were upset were rooted in highly valid concerns; they were about someone's artwork being used without credit to create something else. Other concerns, also with a lot of merit, were shared but I will not go into them here. 

Without a doubt, I agreed with those concerns and the reasoning behind them. 

But, the thought crossed my mind that it was highly likely that others (like me) who struggled to see AI, might post a picture and then find themselves banned because it turned out to have been generated by AI. Banned even if that someone (like me), actually opposed the use of AI and posted it unknowingly.

The model horse hobby is about learning, and there are so many facets within the hobby for that: collecting, painting, tack making, showing, sculpting, and more. We may know something about a few of those facets, but no one is an expert on them all.

My hope is that when something like this happens again (and no doubt it will), we work to educate the person. Perhaps she/he actually agrees about the valid reasons for not using AI and is embarrassed, even mortified, that they actually posted a picture that had been generated by it.

She/he may be (like me) unable to tell it's AI, but would be very happy if folks reached out and tried to help them learn how to see it. 

I can't tell AI generated pictures yet. But that's the key word here - yet. 

Just a thought from me.