artificial-intelligence – Leonberger Life

I Promise Not to Post AI Generated Comments

Esther’s writing prompt: December 3 : Promise

This post is not about Leonbergers but about a phenomenon that seems to be spreading like wildfire in the blogosphere, AI generated comments. If you are a blogger that has been blogging lately, I am sure you have seen them. Well written, seemingly creative, flattering and detailed comments that give the impression that the commentor, who appears to be an extremely articulate writer, has read your post really well and understood it, and has spent a long time perfecting the comment. Unfortunately, he/she didn’t write anything and has probably not even looked at your post. He/she might have, in a very short time, generated hundreds of comments automatically for hundreds of blogs he/she has never looked at and knows nothing about.

When I started getting the AI generated comments I was really flattered at first. Someone really took the time to read my often lengthy and complicated posts and not only understood them well but loved them for all the right reasons. As the Devil in the Devil’s advocate said, “vanity is my favorite sin”. However, I eventually realized that these comments were inauthentic. I still say thank you when I get an elaborate and flattering comment that does not look genuine or seems to be too good to be true. First of all, it is difficult to know for sure whether the comment is AI-generated and secondly, they did make some effort even though it was minimum. I don’t know if that is right or wrong of me.

To generate an AI-comment you can, for example, use a website such as Postiz. You copy and paste the content of a blog post into the edit box “Original Post Content”, then select the tone, for example, “friendly”, and then you press the button. Then copy and paste the comment in the blog.

There are many other tools for generating comments, blog posts, and other content including HyperwriteAI, which also can make the content sound genuine using your writing-persona. It trains the AI to mimic your unique writing style (you provide examples) ensuring the output sounds like you.

My Murder Moose Example

I am not going to post examples without the consent of the non-authors, so I am giving my own example based on my recent post “What are your Favorite Animals Not Murder Moose Anymore“. I used Postiz, which generated the following comment for me.

Wow, that’s such a cool journey from the Murder Moose to loving dogs!

I love how you turned your childhood memories into a drawing.

And writing a book about Leonbergers is amazing!

Dogs really do bring so much joy and companionship.

Rollo sounds like a great buddy too!

Below is the drawing that Postiz is referring to:

A moose with big antlers and giant sharp teeth. — The murder moose in my second grade story. The picture was generated by ChatGPT.

I posted an AI-generated comment above, which makes it seem like I already broken my promise not to post AI generated comments. However, the comment was not posted in anyone’s comment section but here in my post for illustration purposes. What I promise is to not post AI generated comments on anyone’s blog post without telling them it is an AI generated comment, unlike a lot of bloggers. The picture above is also AI generated but I am stating that it is.

In my opinion, leaving AI generated comments on people’s posts without telling them it is AI generated is dishonest and the same is true when someone posts an AI generated blog post without mentioning that it is AI generated.

The Death of Internet

AI generated comments, bots posting comments, bots arguing with people in the comment sections, bots arguing with other bots in the comment sections, and bots “clicking” massive amounts of reactions (like, love, angry, laughter, etc.). Not to mention the spread of deep fakes These are phenomena spreading like wildfire.

Not all AI generated comments and comments by bots are innocent flattery. Bots frequently spread false information, false narratives, division and anger. AI generated comments and bots are ruining the comment sections in various ways. Before you get into a lengthy argument with someone on internet, here is some expert help to determine whether you are arguing with a bot, or not. On my second blog I posted about the fact that “more than half of Internet Traffic is Bots”. Most of the bot traffic is malicious by the way.

A picture of giant computer screen with a red warning triangle containing an exclamation mark. Underneath the triangle is the text “System Hacked” — Warning of a system hacked. Virus, cyber-attack, malware concept. Asset id: 1916985977 by Sashkin

This makes me wonder if the internet is dying. With that I am not referring to the Dead Internet Conspiracy Theory that claims that Internet is already dead and entirely governed by bots that are controlling us. I am referring to something more realistic, that our internet experience is ruined first by people using AI generated content (comments, posts, etc.) without telling readers, and secondly by bots. Is Internet losing its authenticity and value?

And to reiterate, I promise not to post AI-generated comments on people’s blogs.

Our Leonberger ChatGPT and Google AI

Today I was having fun with ChatGPT and Google AI. I asked ChatGPT the following question : Was there a Leonberger called Bronco?

The answer is shown below

The picture is a screenshot. The ChatGPT answer is below. — I took a screenshot of the answer.

If the text is hard to read in the picture, this is the answer ChatGPT gave.

Yes, there was a Leonberger named Bronco. Bronco, officially named Le Bronco von der Löwenhöhle, lived with the Wikman family in Dallas, Texas, from 2007 until his passing in 2020. Notably, he lived an unusually long life for his breed, reaching almost thirteen years, surpassing the typical life expectancy of Leonbergers, which is around nine years. His story is detailed in the book “The Life and Times of Le Bronco von der Löwenhöhle” by Thomas Wikman, which shares anecdotes of Bronco’s adventures and provides insights into caring for Leonbergers.

I think that sounds right.

Confusing ChatGPT

Next, I asked ChatGPT the following question : Did the Leonberger Le Bronco von der Löwenhöhle rescue Hamsters?

I can add that Le Bronco von der Löwenhöhle is a unique Leonberger name. There is a Leonberger database called the Worldwide Independent Leonberger Database featuring more than 172,000 Leonbergers. It covers most Leonbergers in the world back until the beginning of the 20th century and only three of them were named Bronco including Le Bronco von der Löwenhöhle, which is the only Leonberger named exactly that. As you can see below ChatGPT is confusing our Leonberger Le Bronco von der Löwenhöhle with another dog.

If the text is hard to read in the picture, this is the answer ChatGPT gave.

Yes, Le Bronco von der Löwenhöhle, a Leonberger, is known for his heroic act of rescuing a group of hamsters! This event gained attention when Le Bronco, who was living in Germany, saved several hamsters from a fire in 2011. The hamsters were trapped in their cages, and thanks to Le Bronco’s quick thinking and bravery, he managed to alert his owners and helped rescue the small animals.

His actions were recognized as a remarkable act of kindness and heroism, showing the intelligence and sensitivity that dogs, particularly the Leonberger breed, can possess.

I’ve noticed that ChatGPT is sometimes wrong. For example, when you ask questions on a topic for which there is sparse information ChatGPT frequently gets the answer wrong. Another way to confuse ChatGPT and get bad answers is to ask vague questions or ask questions that cannot be answered precisely. ChatGPT relies on internet data and may base its answers on people’s biases and cliches rather than accurate information.

Google AI on Bronco

Interestingly enough, Google AI gave more extensive and interesting answers to my questions, and unlike ChatGPT it answered both of them correctly and provided links to the sources.

First, I asked : Was there a Leonberger called Bronco?

This is what I got :

The picture is a screenshot. The Google AI answer is below. — Screenshot of answer.

If the text is hard to read in the picture, this is what the screenshot says.

Yes, there was a well-known Leonberger named “Bronco” (Le Bronco von der Löwenhöhle) who lived with the Wikman family in Dallas, Texas, and was the subject of a book and blog called “Leonberger Life”.

Here’s a more detailed look at Bronco and his story:

As I scrolled down there was more

If the text is hard to read in the picture, this is what the screenshot says.

Owner:

Thomas Wikman, a retired automation, robotics, and software engineer, and his wife raised three children, and later lived with their two dogs—a pug and a miniature Australian shepherd.

Book and Blog:

Thomas Wikman wrote a book, “The Life and Times of Le Bronco von der Löwenhöhle” and a blog, “Leonberger Life” about his Leonberger, Bronco.

Longevity:

Bronco lived a relatively long life for a Leonberger, receiving the “Grey Muzzle Award” for living longer than 12 years.

Notable Incidents:

Bronco was known for his bravery and protective nature, saving Daisy’s life and potentially Baylor’s life, as well as scaring off a trespasser.

Death:

Bronco passed away in 2020.

On the right side there are references to two of my blog posts and one post by Annika Perry.

If the text is hard to read in the picture, this is what the screenshot says.

Thomas Wikman had Bronco, a 167-pound Leonberger who, by all accounts, was an extraordinary dog with a remarkable owner.

This is followed by a link to the People Newspapers a Newspaper that published a review of my book.

How did Google AI do on the second question?

I asked Google AI the same question : Did the Leonberger Le Bronco von der Löwenhöhle rescue Hamsters?

If the text is hard to read in the picture, this is what the screenshot says.

Yes, the Leonberger, Le Bronco von der Löwenhöhle, is known for a story where he “rescued” two escaped hamsters by carrying them in his cheeks, though they were later revived by Claudia, Thomas’s wife, performing CPR.

Here’s a more detailed account:

If the text is hard to read in the picture, this is what the screenshot says.

The Hamster Escape Artist:

The kids had pet hamsters named Moldova and Montenegro, who occasionally escaped their cages.

Bronco’s Unique Skill:

Claudia would tell Bronco, “Bronco, find the hamsters,” and he would sniff around the house until he located them.

Examples of Bronco’s “Rescues”:

One time he found them in the linen closet.

Another time he found them on a shelf in the living room.

On one occasion, a friend of David trusted them with his two hamsters, and Bronco later “rescued” them after Claudia noticed they were missing from their cage, with Bronco’s cheeks looking puffy.

The “Rescue” Incident:

Claudia noticed Bronco’s cheeks were puffy, and when she said “Bronco, drop it!”, two hamsters, both unconscious, fell out.

CPR Attempt:

Claudia panicked and performed CPR on the unconscious hamsters, and one hamster revived right away.

Second Hamster’s Revival:

The CPR didn’t seem to work on the other hamster, but Claudia put both back in their cage, and soon the second hamster also woke up.

Keeping the Secret:

They decided to keep the incident to themselves, as “hamsters don’t squeal”.

Google Search for Leonberger

Finally, I did another experiment. I typed in “Leonberger” in Google Search and my blog came up on the first page following the Wikipedia entry for Leonberger, the “People also ask” section, the American Kennel Club web page on Leonbergers, the Things to Know section and the What people say section and then comes my Leonberger Life blog followed by 7 more sites on the first page. Interestingly my blog is ahead of PetMD, the Leonberger Club of America and the United Kennel Club.

The screenshot shows the entry for Leonberger Life … The Origins of the Leonberger Dog …. The Leonberger breed was originally created by — The screenshot shows the Google result for “Leonberger”

In conclusion, I was not that impressed by ChatGPT but I was impressed by Google AI and thought the Google search of Leonberger was fun to see.

Additon : A blogger friend of mine, Miriam Hurdle found the following information, which explains the result. Thank you so much Miriam.

Google AI has a larger database of all the information out there. Probably that’s the reason it has the details of your book and your blog. BBasically Google indicates that ChatGPT is good at idea generation and writing, etc. I think that’s why it doesn’t know the details of your book.

Robotics Research

Someone suggested that I write a little bit about my background in Robotics and the fact that I did a PhD in Robotics (Department of Electrical Engineering and Applied Physics). My thesis involved the emulation of basic animal behavior such as reflexes and fixed action patterns, as well as tree-search based artificial intelligence. I used neural networks in other contexts, but not for my PhD research. It also involved quite a bit of non-linear robot system control and the study of chaotic systems. I implemented my research on many kinds of robots but the robot I used the most was the 7-degrees of freedom Robotics Research Corporation Robot pictured below.

Seven jointed metallic robot surrounded by objects — Robotics Research Corporation Robot

It looks messy in the picture, and I should probably explain. I had hung a few boxes from the ceiling and placed a coatrack in front of the robot. There is also another robot just outside the picture to the left. I placed a sonar sensor, or perhaps more correctly eco-location equipment on the tip of the robot. I used the sonar/eco-location to create a 3D map of the robot’s surroundings. Since the sonar-sensor could not see everything at once I used the arm of the robot to search and investigate the surroundings, including discovering that the boxes were open on one side. The motion of the robot was guided by the sonar and was not directly programmed by me. The robot was able to avoid colliding with the objects, including the other robot as it was moving. I moved objects around and added clothes and hats to the coatrack to make it harder. It looked like an animal figuring out what was around it. This was in 1993.

A diagram showing the possible joint rotations of the RRC robot — The 7 possible joint rotations (degrees of freedom) of the RRC robot.

Maybe I should explain what 7-degrees of freedom mean. One degree of freedom could be a rotation (around a joint) or a translation. To arbitrarily position an object in space you need at least three degrees of freedom. To arbitrarily orient an object in space you need an additional three degrees of freedom. So, to arbitrarily position and orient a tool in space you need at least six degrees of freedom. Industrial robots frequently have six degrees of freedom. However, if you also want the robot arm to avoid collisions with objects in its space you need an additional degree of freedom. That’s why I needed a 7-degrees of freedom robot.

In the picture below an arc welding robot PW-10 is moving around unaware of its surroundings. The RRC robot detects that the PW-10 is in its space and perhaps approaching and quickly gets out of the way. Reflex control in action.

PW-10 Arc Welding Robot to the left. RRC robot on the right trying to avoid collision.

Robotics can be Math heavy (Algebra, Trigonometry, Calculus)

Part of the job of creating the control system is to create the kinematics, dynamics, and perhaps inverse kinematics, for the robot, and this is math heavy. Lots of algebra, calculus and trigonometry. I have included the Jacobian matrixes for the first four degrees of freedom for the RRC robot below. Just to illustrate how math heavy it is, and this is just four degrees of freedom. Note, doing this was not part of my research. It was just a basic mandatory task to get the robot to move predictably before you even get started.

I should say that as soon as I finished my PhD, I was offered a job at JPL in Los Angeles to work on an 11-degree of freedom robot that was supposed to be on the International Space Station. However, the government cut the funding for the project, so I never got to do it. Instead, I took a job with ABB Robotics. I did not get to do the math for the 11-degree robot but one of my first tasks at ABB Robotics in Wisconsin was to create the kinematic matrixes for more than 20 ABB robots.

Algebra is indeed a good skill in many professions. A lot of engineers, scientists, mathematicians, IT professionals use algebra in their work, and algebra teaches you how to think logically and methodically. Therefore, schools should teach it. Which is why I don’t find this common joke funny “Another Day Has Passed And I Didn’t Use Algebra Once”. I don’t find “Another Day Has Passed And I Didn’t Play An Musical Instrument Once” funny either but somehow I think the problem with the joke is more obvious to some people in the latter case.