54 Comments
User's avatar
⭠ Return to thread
Based Florida Man's avatar

My wife writes blog articles for an organization and has ChatGPT (4.0) help review the content before she posts. The AI is really good about suggesting improvements and sometimes she totally rewrites them based on the feedback.

Eventually you could see AI bots having their own blogs and followings. Like if Jeff had an assistant AI bot that could harvest articles and content of the day and suggest things for tomorrow's post. Maybe it could even write the whole post in some cases.

And we, the readers could have AI agents that login and engage on our behalf, using our past posts and other sources to have some unique personality and perspectives.

That would be pretty crazy for a substack to be written all by AI and all the commenters are AI.

I expect by the end of the year someone will something like that working.

Expand full comment
John Cougar Misanthrope's avatar

Your wife may want to read the ChatGPT T&C's very closely before sharing any intellectual property content with the algorithm. Nothing in life is truly free. I vaguely recall an issue with Google Translate and Google claiming 'ownership' of content.

Expand full comment
Based Florida Man's avatar

In her case, she's fine with the data becoming public domain, but it's certainly a good idea to make sure these tools don't own your content simply by using them.

Expand full comment
RunningLogic's avatar

Same with Deep L. They say the premium service doesn’t use the information you post but the free one definitely can. And I don’t think I’d trust their word on the premium version either.

Expand full comment
John Cougar Misanthrope's avatar

It's critical for end-users to read and understand the privacy notice and the terms and conditions to understand how their data is being processed, stored, shared, and potentially sold. People would be shocked at how companies openly monetize their personal data.

Most people don't read them which is reasonable because it's shameless legalese, but they should. Companies like Meta get in trouble, particularly in the EU due to the GDPR, because they do not practice what they promise.

Expand full comment
RunningLogic's avatar

I always do, or at least skim them because there are parts that are pretty boilerplate and parts that don’t concern me (residents of California and stuff like that). But I admit I don’t always understand it all very well.

Expand full comment
HUMDEEDEE's avatar

I wonder if you copied and pasted the TOC into ChatGPT and asked it to summarize how the data is being processed, stored, shared and sold, if it would accurately do so. I've tried to read those damned things, and instantly fall asleep - best sleeping aid there is!

Expand full comment
RebeccaGrrrl's avatar

And remember that Chat can also lie, or provide 'improvised' material. I forget how it works, but you have to tell it in the prompt how to do it and avoid generating false information.

Expand full comment
RunningLogic's avatar

Interesting idea! I’d like to see what it would come up with.

Yeah maybe we should all keep a copy on our bedside tables to read in case of insomnia 😆😬

Expand full comment
Johnny Be Real's avatar

You are correct that G Translate T&C said anything you put in is theirs to do with as they choose. Our business could not use it because of that.

Expand full comment
ICI Grief (The Rebel's Hike)'s avatar

I heard there are pastors using ChatGPT to write their sermons. Bad idea.

Expand full comment
Faith&FitnessMama's avatar

The Babylon Bee did a whole segment on their podcast giving AI Chat Bot prompts to write sermons. It was pretty hilarious....none of them were good sermons. 😂

Expand full comment
RunningLogic's avatar

Like when someone here (BFM? I can’t recall) asked about Jeff and it made up all this stuff about him being a coffee expert and connaisseur (which may actually be true, I don’t know, he’s certainly a man of many talents 😁) and a whole bunch of other nonsense that all of who have been following Jeff for a long time immediately would know is bunk.

Expand full comment
AJF's avatar

Hmmmmm what does this mean for HS teachers and college professors? Is there a way to identify something written by AI?

Expand full comment
🌱Nard🙏's avatar

There are sites that can check and let you know how likely it is that student content is AI generated. How reliable they are is hard to tell. I tested a few, and they detected suspect content pretty accurately. However, once I edited the content and used my own “voice,” the generators did not detect AI. It’s going to be interesting to see how this plays out in the classroom. As the Head of School in a Charter School, I’m working w my English dept. to try and find a solution. Our kids may have to start writing all of their papers in class. The flipped classroom model may indeed make a comeback.

Expand full comment
Mary Lynn Mortimer's avatar

My son had a class at UF this past spring where the professor made it pen and paper in class just for this reason. My son actually liked it because there was no homework outside of class.

Expand full comment
RunningLogic's avatar

That’s great!

Expand full comment
🌱Nard🙏's avatar

When I was teaching, my kids wrote all of their papers in class. Pen and paper. Computers only for final drafts. Saves a lot of headaches and makes the kids THINK!

Expand full comment
RunningLogic's avatar

Yup writing by hand the old fashioned way might be the new thing 😁 I actually think that would be an improvement, for many reasons.

Expand full comment
nik's avatar

Very little modification is needed, and it is passing the plagiarism checker at our school.

Expand full comment
🌱Nard🙏's avatar

There are new ones developed specially for AI. It’s easy to pass a plagiarism check as not much is plagiarized...just written by “someone” else 🤷‍♀️.

Expand full comment
Katrina the Hurricane's avatar

Wow, as a former teacher, I like not only the “in-class” exercise, but also the pen and paper aspect since most kids have lost the skill and the art of handwriting. As a con, though, it sure does limit the editing aspect to yield a great paper.

Expand full comment
Susan In Kansas's avatar

Teachers can still hand kids a pen and paper. Writing can be done in the presence of the teacher.

Expand full comment
AJ#2's avatar

Oral exams like in the olden days.

Expand full comment
Oma's avatar

I asked my second year college student grandson if AI was used by anyone he knew. He said casually, “I’d say 6 out of 10 at school use it. Some to do their work and some use it to ‘check’ their work.” I was amazed at that number! My brother (a retired military “original” computer language programmer) and I played around with AI and amazed, frightened and angered at what it can do! As far as how can you tell if your students are using AI, on the first day of class, and every day for a week, pull out pencils and papers that YOU have and tell them to write about what will give you a ‘picture’ of their abilities through vocabulary, sentence structure, thoughts and sometimes even attitudes and personality. That’s what my brother and I came to the conclusion of. Just a thought for you, AJ.

Expand full comment
SeeingTruth's avatar

If teachers/professors write customized questions, assignments, critical thinking exercises, etc it can help negate the issue. It has worked for my spouse. And, the teachers who take the time to know their students can see through the bs.

Expand full comment
AJF's avatar

Asking as a retired teacher 😳

Expand full comment
Oma's avatar

Maybe become a teacher’s assistant and help in an area that this would be a issue. We need to not leave this area of need, but become part of the solution. I know...easier said than done. I worked 27 years in elementary through high school (yes, I was following my 3 children, guilty as charged by them) and I just could not make those last 3 years. And that was the late 70’s through July 31, 1999! My daughter is in her 27th year of ESE elementary through high school and her daughter will begin her career teaching 3rd grade this fall. Not where I wanted her to go but I’m going to help where I can while my daughter recoups from 30 years of Hearing Impaired teaching/supervisor while having an adopted deaf child (political incorrectness changed again- not deaf,hearing impaired. Not hearing impaired, deaf)! Politics were in education in the 70’s. I just didn’t know how deeply it was planted.

Expand full comment
nik's avatar

Already happening at the HS level …and no detection.

Expand full comment
AJF's avatar

🫤

Expand full comment
Help Needed in KS's avatar

BFM, how do you know it's not happening now?

Expand full comment
Based Florida Man's avatar

Wouldn't that be interesting.

Expand full comment
Janet's avatar

Ugh. I’m just about ready to become a hermit.

Expand full comment
RunningLogic's avatar

We need our own C&C country!

Expand full comment
none none's avatar

I really wish there was some way to create C&C groups, maybe by state, and meet once a month for discussion. Wouldn't that be fun!?

Expand full comment
RunningLogic's avatar

It would be great!

Expand full comment
Roger Beal's avatar

Oh, how awesomely cool. Soon we can get rid of pesky human creatives altogether!

Expand full comment
GreaterIsrahell's avatar

"And we, the readers could have AI agents that login and engage on our behalf, using our past posts and other sources to have some unique personality and perspectives."

Yeah, Jeff could be a dumbed down TV watching couch potato, and so could all of us. Wouldn't that be wonderful? LOL

Expand full comment
RJ Rambler's avatar

WALL-E

Expand full comment
Roger Beal's avatar

And who would be the customer for such an arrangement? The Big Robot On Mars?

I cannot imagine what sort of actual thinking human would spend ten nanoseconds to learn what one AI bot says to another AI bot.

Expand full comment
Based Florida Man's avatar

Most substack (including C&C) readers do not leave comments.

So if an AI bot created Tuesday's post, and then a bot that had used my personality (from harvesting my previous posts, and access to other personal sources) would create a comment that would sound roughly like what I would say. And similarly a bot could do the same for you. A thread would emerge like the thread of today, with the bots for each of us posting comments that could be pretty interesting and relevant.

So the customer would be the same customer that reads these substacks, but doesn't post; they would just continue to read the content, assuming it's interesting, if it's AI generated or not.

Pretty wild, huh!

Expand full comment
Roger Beal's avatar

Wild ... and ultimately pointless. Why spend time reading the thoughts of a "real" bot followed by my comments as simulated by a "fake" bot?

Perusing cereal boxes over breakfast is a better use of time, IMO. (:>)

Expand full comment
Based Florida Man's avatar

Simply refer to Jeff's post today and how Ms. Marsh is fake with an AI generated photo. And hundreds of thousands of followers and retweeters.

Fake content, with millions of views!

I mean, any video game has Non Playable Characters (bots) that engage hours of people's time daily. They pay big money for fully simulated content.

So why wouldn't people also read blogs with comments that are all AI generated, assuming it was interesting?

Expand full comment
Fla Mom's avatar

And how would they know it's all AI-generated to begin with? They wouldn't.

Expand full comment
RunningLogic's avatar

How do we know all of her “followers” (or even a majority) aren’t also bots?

Expand full comment
none none's avatar

The bottom line is, we can assume that very soon (now) we won't be able to tell the difference between a bot and a real person. At least online.

Expand full comment
Tamsin's avatar

I thought you were... based. 🤨 It's right there in the name.

Expand full comment
Based Florida Man's avatar

I'm just exploring the topic of AI that Jeff brought up today.

Obviously we'll be seeing much more AI generated material, so we're just thinking aloud.

Expand full comment
RJ Rambler's avatar

Actually, good point because it's JUST LIKE THAT!

Expand full comment
daverkb's avatar

It would be a Transhumanist Desert operating as a Mirage Oasis. Death Incarnate.

Expand full comment
Shelle's avatar

Jeff has a unique writing voice and it would be so sad to have that replaced with AI! Or to have AI mimicking it in some "off" way

Expand full comment
User's avatar
Comment deleted
Jul 3, 2023
Comment deleted
Expand full comment
RunningLogic's avatar

It sounds like a pretentious middle school or high school student.

Expand full comment