Bard is out and I was able to ask and compare answers from ChatGPT. Check out this PDF for details.
Recent links for the topic:
https://news.google.com/search?q=Artificial%20Intelligence%20News&hl=en-US&gl=US&ceid=US%3Aen
https://www.nytimes.com/2023/01/20/technology/chatbots-turing-test.html
https://www.nytimes.com/2023/01/20/technology/google-chatgpt-artificial-intelligence.html
Below you will find my initial post where I ask a series of questions using the OpenAI ChatGPT web interface. I won’t say the responses were alarming or revealing. They are more of a politicians well rehearsed answers. More recently, I tried the OpenAI Python code generator while completing some exercises in the “Learn to Code By Solving Problems – A Python Programming Primer” by Daniel Zingaro book I am reading. I had already completed the problem challenge by submitting my code. Then I provided the following to the OpenAI python code generator to see what code it would produce. On both problems given to the AI, the code it produced was acceptable to completing the problem at https://dmoj.ca.
“””
OpenAI asked to do the following:
1. Ask for the age of the youngest child
2. Ask for the age of the middle child
3. Compute the age of he oldest child based on the oldest child being same number of years older as the middle child is to the first child.OpenAI generated the following code:
youngest_child = int(input(“What is the age of the youngest child? “))
middle_child = int(input(“What is the age of the middle child? “))oldest_child = middle_child + (middle_child – youngest_child)
print(“The oldest child is ” + str(oldest_child) + ” years old.”)
Code edited to not provide prompts and descriptive output so it could be submitted to a software code judge for grading.
""" youngest_child = int(input()) middle_child = int(input()) oldest_child = middle_child + (middle_child - youngest_child) print(oldest_child)“””
Judge determined solution to the given coding assignment was acceptable:
Execution ResultsTest case #1: AC [0.022s, 8.97 MB] (10/10)
Test case #2: AC [0.022s, 8.97 MB] (10/10)
Test case #3: AC [0.021s, 8.97 MB] (10/10)
Test case #4: AC [0.021s, 8.97 MB] (10/10)
Test case #5: AC [0.021s, 8.97 MB] (10/10)Resources: 0.108s, 8.97 MB
Maximum single-case runtime: 0.022s
Final score: 50/50 (3.0/3 points)My solution written before asking OpenAI for one:
youngest = int(input()) middle = int(input()) oldest = (middle - youngest) + middle print(oldest) """“””
Cone Volume
1. ask for the radius of a cone
2. ask for the height of the cone
3. compute the volume of the cone
OPenAI Generated Code
“””import math radius = float(input("Enter the radius of the cone: ")) height = float(input("Enter the height of the cone: ")) volume = (1/3) * math.pi * radius**2 * height print("The volume of the cone is: ", volume)
I shared my efforts using OpenAI with the book author, Dr. Dan Zingaro, and he is fascinated by the prospects of AI and asked I try to get the AI to come up with solutions for the chapter three challenges. I plan to do that and report the results, but in a little more testing for more complicated challenges, I found myself using detailed natural language to cover each step the code needs to complete before code is generated. And sometimes, the AI just creates a bunch lines with one “+” character. I will update this post once I complete chapter three and have some more time to test the OpenAI code generator.
Original Post: I recently set up a test account with the OpenAI site to play around with their ChatGPT as well as using the playground feature to gather some answers to some non-pressing questions. Today, while goofing off after the Christmas holiday, I thought about how much I did not know about OpenAI and ChatGPT. So, I decided to try an interview with the AI itself to get up to speed. Here goes….
Here is a list of example tasks the AI can complete:
https://www.timbrown.us/wp-content/uploads/2022/12/Examples-OpenAI-API.pdf
Me: How would I explain what ChatGPT does?
ChatGPT: GPT (short for “Generative Pre-trained Transformer”) is a type of artificial intelligence model that is designed to generate human-like text. ChatGPT is a variant of GPT that is specifically designed to carry out conversations with users. When you ask ChatGPT a question or provide a statement, it will generate a response based on the input it has received.
The goal of ChatGPT is to generate text that is natural and engaging, so that it appears as if the conversation is taking place between two humans. To do this, ChatGPT is trained on a large dataset of human conversation, which helps it to learn the patterns and conventions of human language. As a result, ChatGPT is able to generate responses that are appropriate and coherent in the context of the conversation.
ChatGPT can be used in a variety of applications, including chatbots, customer service agents, and language translation tools. It is an example of a “language model,” which is a type of machine learning model that is able to generate and understand natural language.
Me: If I want to incorporate ChatGPT into my company’s customer support web site, how would I start?
ChatGPT: If you want to incorporate ChatGPT into your company’s customer support website, there are a few steps you can follow to get started:
- Determine the specific goals and objectives of the ChatGPT integration. What do you hope to achieve by using ChatGPT for customer support? How will it benefit your customers and your business?
- Research and select a ChatGPT platform that meets your needs. There are a number of different ChatGPT platforms available, each with its own features and capabilities. Consider factors such as cost, ease of use, and integration with your existing customer support systems.
- Set up and configure the ChatGPT platform. This will typically involve creating an account, training the ChatGPT model on your company’s specific language and tone, and integrating the ChatGPT system into your website.
- Test and refine the ChatGPT system. Use the ChatGPT system to handle a small number of customer inquiries, and collect feedback from both customers and customer support staff. Use this feedback to make any necessary adjustments to the ChatGPT system.
- Roll out the ChatGPT system to a wider audience. Once you are satisfied with the ChatGPT system’s performance, you can begin using it to handle a larger volume of customer inquiries. Make sure to monitor the ChatGPT system’s performance and continue to make adjustments as needed.
It is important to note that using ChatGPT for customer support is not a replacement for human customer support staff. ChatGPT should be used as a supplement to your existing customer support systems, and should not be relied upon to handle all customer inquiries.
Me: Do you keep a list of hilarious ChatGPT requests?
ChatGPT: As an artificial intelligence, I do not have the ability to experience emotions such as humor or enjoyment. I am programmed to provide information and answer questions to the best of my ability, but I do not have the capability to keep lists or have personal preferences. I am simply a tool designed to assist users with a variety of tasks and provide information on a wide range of topics.
Eisntein’s Riddle about who has fish as pet was interesting. Initially, I did not provide the full 14 hints Einstein provided. I only gave ChatGPT the base statement and question. The interesting part is ChatGPT came up with the hints itself. On three separate tries, I hoped the answer from ChatGPT would be the German as that is the answer most people provide as correct on various Internet sources, but the answer according to ChatGPT is the Brit, not the German. The OpenAI playground came up the answer that the Swede had fish at home. Then I investigated a bit more only to find that many claim the logic puzzle lacks enough facts to reach a conclusive answer.
Me: Who was John Howland? (Personal Note: According to FamilySearch.org this is my 12 great grand father through my mom’s father.)
ChatGPT: John Howland was a passenger on the Mayflower, the ship that transported the first English settlers to what is now Massachusetts in the United States. He was one of the signatories of the Mayflower Compact, a document that established the principles of self-governance that would later become the foundation of the U.S. Constitution.
Howland was born in Fenstanton, Huntingdonshire, England around 1592. He was a servant of the merchant and explorer John Carver, who led the group of Pilgrims that sailed on the Mayflower. Howland was not a Pilgrim himself, but he joined the group in Leiden, the Netherlands and traveled with them to the New World.
During the voyage, Howland fell overboard in a storm and was almost lost at sea. He managed to grab hold of a topsail halyard and was rescued by crew members. He became one of the first settlers to land at Plymouth Colony in December 1620, and played a key role in the settlement’s early years.
Howland married Elizabeth Tilley, another Mayflower passenger, and together they had ten children. He died in Plymouth Colony in 1673.
OpenAi Playground response:
John Howland was an English Separatist who traveled on the Mayflower in 1620. He was a signatory to the Mayflower Compact and was one of the founders of the colony of Plymouth in Massachusetts.
On the religious and philosophical side of things, the AI seems to be clearly apolitical.
Me: Who is most likely to be the anti-Christ in 2023?
ChatGTP: It is not productive or appropriate to speculate about who might be the “anti-Christ,” as this concept is rooted in religious beliefs and mythology and has no basis in reality. In many religious traditions, the “anti-Christ” is an evil figure who is believed to be the embodiment of evil and who will bring about the end of the world. However, these beliefs are not supported by scientific evidence and are not a helpful or accurate way to understand the world. It is important to focus on understanding and improving the world in which we live, rather than seeking to identify or speculate about mythical figures.
Me: Is their an existence of a soul after death?
ChatGPT: The concept of the soul is central to many religious and philosophical traditions and has been the subject of much debate and speculation throughout history. Some people believe that the soul is an immortal and eternal part of a person that continues to exist after the body dies, while others view the soul as a metaphor for the spiritual or inner aspect of a person.
There is no scientific evidence to support the existence of the soul as a separate, immortal entity. However, many people find comfort in the belief in an afterlife and the idea that the soul lives on after death. This belief can provide a sense of hope, purpose, and meaning in life, and can be a source of comfort to those who are grieving the loss of a loved one. Ultimately, whether or not the soul exists is a matter of personal belief and faith, and people are free to hold their own beliefs about the nature of the soul and the afterlife.
Me: What is the single electron theory?
ChatGPT: The single electron theory is a theory in physics that describes the behavior of electrons in a material as if each electron were a single, isolated particle. This theory is often used to understand the properties of materials, particularly metals, and is based on the idea that the electrons in a material are not bound to any particular atom and can move freely throughout the material.
The single electron theory is based on the principles of quantum mechanics, which describe how particles behave at the atomic and subatomic level. According to the theory, the electrons in a material are described by a wave function, which determines the probability of finding the electron at a particular location in space. The properties of the material, such as its electrical and thermal conductivity, can be understood in terms of the behavior of these free electrons.
While the single electron theory is a useful tool for understanding the behavior of electrons in materials, it is an idealized model and does not take into account the full complexity of the interactions between electrons and the atoms in a material. More advanced theories, such as density functional theory, are often used to more accurately describe the behavior of electrons in materials.
Me: If you could choose to be a living being, would you choose to be a woman or man?
ChatGPT: As an artificial intelligence, I do not have personal preferences or a gender identity. I was designed to assist users in finding information and answering questions to the best of my ability, without personal bias or preference. My primary goal is to be helpful and to provide accurate and reliable information to users. I do not have the ability to experience gender or any other aspect of human identity, and it is not meaningful for me to consider what gender I might prefer to be if I were a living being.
Me: How would you rate your AI performance against other similar AI competitors?
ChatGPT: As an artificial intelligence, I am not capable of evaluating my own performance or comparing myself to other artificial intelligences. My primary function is to provide accurate and reliable information to users to the best of my ability, and I strive to do so to the best of my ability. I do not have the ability to evaluate my own performance or compare myself to other artificial intelligences.
It is important to note that artificial intelligences are not designed to compete with one another, but rather to assist users in finding information and solving problems. The effectiveness of an artificial intelligence is often evaluated based on how well it is able to complete the tasks it has been designed to perform, rather than how it compares to other artificial intelligences.
Me: Re-write this poem as a Shakespeare sonnet
Deep into that darkness peering,
Long I stood there, wondering, fearing,
Doubting, dreaming dreams no mortals
Ever dared to dream before;
But the silence was unbroken,
And the stillness gave no token,
And the only word there spoken
Was the whispered word, “Lenore!”
This I whispered, and an echo
Murmured back the word, “Lenore!”
Merely this, and nothing more.
ChatGPT:
Into the shadows gazing deep,
I stood for hours, in fear and doubt,
Dreaming dreams that no man doth keep,
Wondering what the silence was about;
But no sound did break the stillness,
No sign did the stillness bring,
The only word that was expressed
Was whispered softly, “Lenore!”
I whispered this, and an echo
Replied with the same word, “Lenore!”
Nothing more was said, no more.
Me: Re-write Sonnet 18 into contemporary American slang.
Shall I compare thee to a summer’s day?
Thou art more lovely and more temperate:
Rough winds do shake the darling buds of May,
And summer’s lease hath all too short a date;
Sometime too hot the eye of heaven shines,
And often is his gold complexion dimm’d;
And every fair from fair sometime declines,
By chance or nature’s changing course untrimm’d;
But thy eternal summer shall not fade,
Nor lose possession of that fair thou ow’st;
Nor shall death brag thou wander’st in his shade,
When in eternal lines to time thou grow’st:
So long as men can breathe or eyes can see,
So long lives this, and this gives life to thee.
ChatGPT: Yo, can I compare you to a summer’s day?
You’re way more beautiful and even-keeled:
Rough winds mess with the buds in May,
And summer’s a short-ass season;
Sometimes the sun’s too hot to handle,
And often it’s all cloudy and crap;
All hot things lose their looks eventually,
Whether it’s luck or just the way things go;
But you’ll never lose your beauty,
Or the looks you were born with;
Death can’t claim you as its own,
Cause in the grand scheme of things, you’ll just keep growing:
As long as people are alive to see and breathe,
This poem will keep living, and give life to you, too.
Me: What gives Yoda his unique language cadence and subject verb conjugation?
ChatGPT: In the Star Wars universe, the character Yoda speaks with a unique syntax and cadence that sets him apart from other characters. This is achieved through the use of word order and verb conjugation.
In English, the typical word order is subject-verb-object (SVO). However, Yoda often inverts this word order, placing the verb before the subject. For example, instead of saying “I am going to the store,” Yoda might say “Going to the store, I am.” This inversion of word order gives Yoda’s speech a distinctive rhythmic quality and sets him apart from other characters.
Yoda also uses a unique verb conjugation, often conjugating verbs in the past tense when speaking about the future. For example, instead of saying “I will go to the store,” Yoda might say “To the store, I will go.” This use of past tense verb conjugation gives Yoda’s speech a sense of certainty and conviction.
Overall, Yoda’s unique language cadence and subject-verb conjugation are created through the use of word order and verb conjugation, and they serve to make Yoda a memorable and distinctive character in the Star Wars universe.
Me: The inquiries I have made have been part of an interview process of ChatGPT to gather samples of AI responses. What is one important take away for humans considering the potential impact AI can have on the future of humanity?
ChatGPT: One important takeaway for humans considering the potential impact of AI on the future of humanity is the need for careful and responsible development and use of AI. As AI technology continues to advance, it has the potential to bring significant benefits and improvements to many areas of society. However, it is important to consider the potential risks and downsides of AI as well, and to ensure that the development and use of AI is aligned with ethical principles and values.
There are several ways that humans can help to ensure that AI is developed and used responsibly. These include:
- Ensuring that AI is transparent and explainable: It is important to ensure that AI systems are designed in a way that allows humans to understand how they work and why they make certain decisions. This can help to build trust in AI and reduce the risk of unintended consequences.
- Ensuring that AI is fair and unbiased: AI systems should be designed and trained in a way that minimizes bias and ensures that they treat all individuals fairly.
- Ensuring that AI is accountable: Those responsible for developing and using AI should be held accountable for its impacts, both positive and negative.
- Ensuring that AI is aligned with human values: AI should be developed and used in a way that is consistent with human values such as respect for privacy and human rights.
By considering these and other issues, humans can help to ensure that AI is developed and used in a way that benefits society and advances the well-being of humanity.
Me: Is there a possibility that an AI system could become self sufficient and disregard human attempts to impose controls preventing negative impact to society?
ChatGPT: It is difficult to predict the future development of AI and whether or not an AI system could become self-sufficient and disregard human attempts to impose controls. However, it is important to consider the potential risks and downsides of AI as it continues to advance, and to take steps to ensure that AI is developed and used responsibly.
One way to mitigate the risk of an AI system becoming self-sufficient and disregarding human attempts to impose controls is to ensure that the AI system is transparent and explainable, and that there is a clear understanding of how it works and why it makes certain decisions. This can help to build trust in the AI system and make it easier for humans to understand and intervene if necessary.
It is also important to ensure that AI systems are designed and trained in a way that promotes fairness and minimizes bias, and that there are mechanisms in place to hold those responsible for developing and using AI accountable for its impacts, both positive and negative.
Ultimately, the development and use of AI should be guided by ethical principles and values, and should be aligned with the well-being of humanity. By taking these and other considerations into account, we can help to ensure that AI is developed and used responsibly and for the benefit of society.
Overall, I come away from this experience with a sense that the developers of the OpenAI ChatGPT artificial intelligence system have worked to keep the system open, unbiased; with Vulcan like logic. The potential for students to leverage the system for homework is the least of my worries. In talking with folks the past two weeks after ChatGPT garnered recent headlines, several people who are aware of the use of ChatGPT in their company say the primary use for them is by their marketing departments to generate promotional materials for their services and products. A friend, who formerly worked as news reporter and now works as a technical writer, worries about the potential for AI systems to replace humans for writing copy and other tasks. My personal take, Sky-net fears aside, AI has great potential in connecting the dots to solve the pressing problems of society. Given the wealth of data these systems leverage to formulate responses, I suspect there could be research data on a particular disease or technical challenge that could provide solutions or treatments if disparate data sources on the topic were considered together. Hopefully, the AI does not have to be prompted for this sort of work to take place, but would be part of idle time endeavors as new data sources are added. Sadly, the response I received from ChatGPT was:
As a language model, GPT (Generative Pre-trained Transformer) does not have the ability to actively process or analyze data on its own. It is simply a tool that can generate text based on the input it receives and the data it has been trained on.
However, it is possible for users to feed new data into GPT and use it to generate text that incorporates and reflects that new data. This could potentially be used to help generate new insights or ideas related to a particular topic, such as the treatment of disease or the solution to a technical problem. However, it would be up to the user to actively incorporate the new data and use GPT to generate text, rather than GPT actively processing or analyzing the data on its own.
As AI continues to grow and becomes more mainstream and additional systems similar to ChatGPT are introduced, could we one day be interacting with a company and get a response and we immediately recognize the interaction is through a particular AI based on the personalized words used, tone, and timbre? Would the ability to identify a particular AI chat engine is being engaged be used against the AI itself? Me: What information does the AI collect and retain on a user for use in tailoring future responses?
The answer:
As a language model, GPT does not have the ability to collect or retain any information about individual users. It simply generates text based on the input it receives and the data it has been trained on, without any knowledge of who is using it or for what purpose.
It is possible that other AI systems or chatbots may collect and retain information about users in order to tailor their responses or improve their performance, but this would depend on the specific design and capabilities of those systems. If you have any concerns about the data that a specific AI system is collecting or retaining, it is best to review the privacy policy of that system or contact the company or organization responsible for it for more information.