Wow, that's scary - not that I know much about Charles de Mills (except it suckers?), so I have no idea if it's correct. Or are you joshing us Brandon, and you wrote this?!

The only thing that I wrote was the title of the post and the paragraph before the essay. there is information that i believe to be incorrect in it

Well, Chat GTP sure goes on and on.... Not the most interesting description of this beautiful rose. Also, for sure incorrect information in there. Isn't it a once bloomer that should prunedvafterprunedprunedprunedvafterprunedprune

Be pruned after it blooms?

Thanks for sharing. That was entertaining.

@librarian_gardner_8b_pnw yes Charles de Mills is a once blooming rose and shouldnt be pruned until after it blooms; though I question pruning it beyond removing dead and diseased canes.

also Charles de mills doesn’t normally produce many, if any, hips on its own. yes i know paul barden used it as a seed parent but that wasnt open pollination that he did.

Brandon -- I've been playing with chat gpt since December -- I've asked it a number of questions. When it can gather common knowledge from the internet, it writes decent essays. Anything requiring a deeper dive, however -- and it literally fabricates references, information, etc. I asked it to tell me about myself -- and it got about 10% right, made up the other 90%, but made it sound totally believable. When I accused the AI of making @#$# up, it said it is AI, and cannot "make things up" but it can be confused and make mistakes, sometimes. So strange. I've warned my students that if they don't know their topic, they could believe the AI is doing a good job, but their professors will see right through it. And that's true if it gets stuff wrong that the professor knows about -- but I think I could easily be fooled by AI on a straightforward exam-style essay...fooled meaning that I wouldn't know if a student wrote it or not. I should ask it to write an article for me...

Lots of charm was mentioned--at least three times. I prefer Wiki. And HAL 9000. Diane

Good Lord. That essay is electrifyingly dull.

Brandon, thanks for doing the dirty work.

Oh my. It's so strange that it just makes stuff up.

The essay is dull and inaccurate. But what happens if you ask ChatGTP to revise the essay to be more interesting and accurate (and, for good measure, concise)?

I added the words ”that is concise and accurate” to the end of the original request. Below is the result. it is far worse..

The Charles de Mills rose is a hybrid perpetual rose that was first introduced in France in the mid-19th century. This rose is known for its deep red, almost maroon, petals that have a velvety texture and a slight fragrance. It is a popular choice among rose enthusiasts due to its striking appearance and its ability to bloom repeatedly throughout the growing season.

The Charles de Mills rose was developed by the French rose breeder Jean-Baptiste Guillot in 1837. Guillot was known for his work in creating new rose varieties, and the Charles de Mills was one of his most successful creations. This rose was named in honor of a wealthy Frenchman named Charles de Mills, who was a patron of the arts and a supporter of rose breeding.

The Charles de Mills rose is a vigorous grower that can reach up to six feet in height and four feet in width. It has a bushy growth habit with thorny stems and dark green foliage. The flowers of this rose can grow up to four inches in diameter and are double-layered with up to 40 petals each. The petals are a deep red color that fades slightly towards the edges, giving the flowers a striking, velvety appearance.

One of the most notable features of the Charles de Mills rose is its ability to bloom repeatedly throughout the growing season. This rose typically blooms in late spring or early summer, and then again in late summer or early fall. This extended blooming period makes it a popular choice for gardeners who want to enjoy the beauty of roses throughout the growing season.

In addition to its striking appearance and extended blooming period, the Charles de Mills rose is also known for its disease resistance. This rose is relatively resistant to common rose diseases, including black spot and powdery mildew, which can be a significant issue for other rose varieties.

When it comes to growing the Charles de Mills rose, it is important to provide it with the right growing conditions. This rose prefers full sun or partial shade and well-drained soil. It is also important to prune this rose regularly to encourage new growth and maintain its shape. Pruning should be done in the early spring, before new growth begins.

Overall, the Charles de Mills rose is a stunning and reliable rose variety that is popular among gardeners and rose enthusiasts alike. With its deep red petals, velvety texture, and extended blooming period, this rose is sure to add a touch of beauty to any garden or landscape

@DDinSB (Z10b Coastal CA) I asked it to write an article about me using my full name and it spit out words that were not related to me. I then gave it more infomation to see if it would give an article that was actually about me. This was its response

I'm sorry, but as an AI language model, I cannot generate an article about an individual without accurate and up-to-date information. Additionally, I cannot browse the internet or access real-time information. As of my knowledge cutoff date of September 2021, there was no publicly available information on Brandon Grant Garner living in Oregon and Georgia or on the passing of his father in an accident.

so if it can’t browse the internet where is it coming up with any of its infomation?

In the past week, we have had 3 students (high school) use Chat GPT to write their required essays. To an adult, the Frankensteinesque quality of the writing is clear, but sadly the students are impressed by its speed and volume. I am going back to handwritten essays done in class only.

Ugh. How can we be so clever (as a species) and yet get dumber every day??

I ran this brief descripton in DallE 2 which is related-ish to Chat GPT to see what would hapen. the pict below is one of the results. Charles de Mills, a gallica rose, mauve in color, matte green foliage, fully double blooms

Here is another

And another

One more

AnAnd thithis is what happened when i just put in ”the rose Charles de Mills”

And these

while nicelooking roses, none of them are Charles de Mills

If you call it "mauve", then your far less likely to get an "accurate" result than if you use the word "crimson".

Charles de Mills, a gallica rose, crimson in color, matte green foliage, fully double blooms

Charles de Mills, a gallica rose, crimscrimson, slate slate-purple in color, matte green foliage, fully double blooms

Looks like I don't have to worry about dumb AI tools displacing my real photos of "Old Chuck"!

While neat, stil not Charles de Mills. i feel like i have seen a picture of a rose similar to the one on the top of this specific post

@User I don’t think that anyone can top your pictures of Charles de Mills. Wh When you had your website up and then when it was being hosted someone else, I would always pull up your page on Charles de Mills when asked what my favorite rose was and show them and tell them it was by you and the best picture that I know of. Heck, I struggle to like any of my own pictures of it because I mentally compare them to yours.

@Brandon Garner Thank you. Unfortunately the photos of 'Charles de Mills' I posted years ago are some of the most often stolen/misappropriated images I've published. I can find my photos of the rose on a variety of rose nurseries around the globe and on personal web sites - often with the original watermark present and with other people's "copyright" on the image! This is one reason why I opted to stop publishing new work and to delete my web site. (File it under: no good deed goes unpunished)

With every passing year, it seems the Web becomes more and more hostile to those who create unique work.

@User sorry to hear that there are that many people with misplaced (or no) morals regarding your work. Especially rose people. I really enjoyed your website, as others I am sure, and speaking for me would rather jut refer people to your website for pictures of roses.

This is slightly off topic, though still about Chat GPT and similar. A comment I read on a news site noted that Chat GPT could do the work of writing for people who don't know how to write; I don't mean pure illiteracy, but people who have a weak command of grammar, spelling, sentence construction, etc. I didn't understand what bothered me particularly about this observation, but have now figured it out: if a person can't write, can that person read? The two skills are linked! What happens then to communication?

This is aside from the obvious problems with Chap GPT, many of them brought to light in this thread.

I must say, I find the whole AI thing deeply, deeply troubling. Of course it has its place (I understand it's a really helpful tool in analyzing some medical tests, for example). But using it to try to replace human creativity is really kind of horrible. I've seen some of those "can you tell the AI art from the human art" tests, and I always get them right. And as a musician, don't get me started on AI music! But sadly, our culture is getting dumbed down enough that our eyes and ears are being trained to accept this stuff as "good enough". End result being, lots of artists and composers (and I guess writers) out of paying gigs, and a gradual erosion of one of the most crucial parts of our souls, the part that responds to truth and beauty and art. A loss of our basic humanity. I love Miyazaki's statement that AI art is "an insult to life itself".

@frances: I'm offering a solution that's probably more a vision than what we're likely to see, but what if children grow up learning to make music, make visual art, write creatively? I have a strong suspicion that people who've learned an art form, learned it well, will be better able to spot AI creations than those who haven't. Exposure to art and music might help, too, but making really sharpens the mind. P.S. This creative activity could also extend to crafts: ceramic, fabrics, woodworking.

Melissa, I think you are spot on @ how learning how to make music, art, written compositions etc. --and sticking to learning until one can do it well enough -- helps a person spot AI. In my classroom, the students who can "do" literacy well have not found AI appealing and can pick it out even if they can't explain why it seems "off". The students who are reading and writing at levels several years behind their peers find AI wonderful.

And, reading is much easier than writing. I like to explain it this way: I can read Shakespeare, but I can't write Shakespeare. Reading is taking in and writing is putting out -- linked, but two different processes.

People who use these new AI tools to make music or create images are learning little-to-nothing about making art or making music, they are only learning how to use a tool that does all the work for them. They are not learning a skill or a craft. I'm not saying that interesting things won't come out of this technology, but what is the user learning by engaging such a tool??

@User I learned by using theses two tools, that the AI system has a long way to go, especially on things that there may not be lots of common knowledge, and I think that I would just prefer to either read a book or come to this forum when it comes to roses than AI

@Brandon Garner - sorry for the delay in answering. Apparently, the AI does a sweep of all Internet content available to it -- at least the stuff not behind pay walls -- as of a certain date, and digests all that. So if you ask it to do a search for you right now, it can't. If you give it more information about you, it just uses what you gave it.

I agree, @Melissa Northern Italy zone 8 -- what happens to communication in this brave new world?? I think AI may eventually help us learn things -- but right now it's just not ready. My friend is a programmer, and he says this will generate code, but you can try it instantly -- and code works or it doesn't. So at least you know if your code is "right" !

Another professor said that if a student is worried the essay will be too "good" the student can just ask for a first draft, and then ask for a second draft with a couple of grammatical errors! Then it's harder to detect.

One thing I find scary about AI is the concept of the self-reinforcing loop. What I mean is, those computer-generated "CdM" images and those bland essays are now online next to the name "Charles de Mills". The next time the AI bot scans the web, this page will be fed into its training data. The next time someone asks about CdM, the bot might regurgitate this page (with a few words thesaurus-ized) as fact! So the truth can get ever more obscured.

A co-worker who was present for the early days of PC use in the engineering trades shared this pithy saying about over-reliance on computer programs and reports: "Garbage in, gospel out."

The results of a computer analysis can only be as good as the data that is fed in. And that is *if* the program itself does everything right. There are many ways for such things to go wrong.

I think it's shameful for students to use these tools, as it will cripple their education. The first encounter with real world problems will expose how little they have learned.

On top of those concerns, you have the possibility of deliberate misinformation. This has already come up a little bit with the prevalence of centralized search and social media sites like Google, Facebook, Twitter, etc. But consider how repressive states could use these tools to spread false narratives, either through direct control or subversion of facts. The old saying, "Believe half of what you see, and none of what you hear" is getting ever more true.