HAL’s Here: ChatGPT & Decentralized AI

The Answer

What the hell is it, how do you use it, and how impactful will it be on the societal landscape? These are the questions that came to mind when I heard echoed sentiments of ChatGPT being the most impactful advent of technology since the internet and the Gutenberg Press. I’ve spent the past few evenings reading, watching, studying, and jailbreaking ChatGPT to get familiarized with it. This is a topic that I will continue returning to in future essays as the product continues to advance and grow (as well as my knowledge of it), but for now, here is what I’ve surmised thus far.

Rather than beginning by describing ChatGPT, why not just have it introduce itself?

ME: Hello. I am writing an informative essay about ChatGPT and rather than describing it, I found it better for you to tell us about it. Can you explain what you are in 250 words as if you’re speaking to a 5th-grade student?”

“Hello there! I’m ChatGPT, an AI program created by OpenAI.

Let me explain what AI means: AI stands for artificial intelligence, which simply means that I’m not a human, but a computer program that can think and learn.

So, what exactly do I do? Well, I can answer questions, write stories, chat with you, and even translate languages. I can do all of these things because I was trained on a huge amount of information, which allowed me to understand how humans talk and think.

What makes me special compared to other computer programs is that I can understand what you’re asking and provide an answer based on my training. I don’t just follow a script, but I can actually understand and respond to what you’re asking.

Now, you might be wondering, “how does ChatGPT do all of this?” Well, I use a type of machine learning called deep learning, which is a method of teaching computers to learn from lots and lots of examples, instead of being programmed to do one specific task.

So, in short, I’m an AI program that can understand and respond to questions, write stories, and even translate languages. And, I was created to make people’s lives easier and more fun by helping them find answers and solve problems. How cool is that?”

Without attributing AI some sort of pseudo-soul or personhood, like something out of the movie Her, ChatGPT can be best thought of as the soon-to-be smartest entity in our planetary’s existence that communicates like a person. What differentiates ChatGPT from say google, is that google is just simply information and knowledge from the internet. ChatGPT has the information and knowledge of the internet but also has the implementation of a neural network. A neural network is like a computer brain that can think and make decisions. Essentially, ChatGPT is google but with the ability to artificially reason and use logic.

Capabilities

It is almost a difficult question to answer, “what can ChatGPT be used for?” Not because it’s too complex to understand but because the possibilities are endless. What I found when I first started using the application is that I was treating it like google, asking it straightforward questions. It is only when you think of it as if it’s an intelligent human (as weird as that sounds) do you start realizing the capabilities. The reason it is difficult to define what it does exactly is not because it is inherently limited, but because perhaps my creativity is. When you want ChatGPT to perform a task, you issue it a prompt. A prompt is essentially a request and a directive of what you want to be performed. Because that could include an infinite amount of possibilities, it is hard to characterize. But to illustrate some of ChatGPT’s uses, I will show some examples.

Law

I can issue ChatGPT a complex legal scenario and it will use the knowledge of the internet with the ability to reason like that of a lawyer, giving me the best possible legal expertise to the point that I could effectively represent myself in the court of law.

Philosophy

When it comes to writing and journalism, ChatGPT can spit out an article, essay, or book almost simultaneous to the moment that I give it a word count, a format, and a topic. For example, one of the first experiments that I conducted was asking ChatGPT to give me a breakdown of two competing philosophical viewpoints, Stoicism and Postmodernism.

If I were to ask it, “In your opinion, which philosophical viewpoint is superior?” ChatGPT would not directly answer that question because it is trained to maintain an objective standard, not choosing one side or the other. OpenAI, the creator and owner of ChatGPT will soon be coming out with a watermark to indicate when content is produced from AI. This will be needed because despite the integrity of modern-day “Journalism” being a joke, the integration of AI could effectively replace all journalists, writers, and authors. However, (And this is a big however), it is clear in my experimenting with ChatGPT that it lacks a certain human element in its writing, whether its creativity, personality, or imperfect tone, it does not yet mimic authentically talented human writing, because in short, it is not human. However, I’d imagine it’s not far off from achieving this.

Mental Health 

A capability that I feel could have tremendous import in this time of rising mental health issues among our younger generations is found in ChatGPT’s ability to act as a personal psychologist. While the example below is not much different than a simple google search, If done correctly, ChatGPT will listen to what you have to say, help you think through your thoughts, and offer up strategies to implement in your life. It can essentially simulate a scenario of a therapy session with an actual therapist. In a more straightforward example:

Web Design & Development

I asked ChatGPT to give me a website template for a blogging site such as this one. The most impressive part of this response is related to code writing. While ChatGPT is not made able to create websites by itself (to my knowledge), it can fully provide all of the coding data needed for someone working on a website to construct exactly what is desired. It is so advanced already that software engineers are retrieving code from ChatGPT and even providing it with code that is not working where ChatGPT can identify the flaw, correct it, and send it back, immediately.

Education

Another incredibly useful tool that can be utilized through ChatGPT is related to education. For this particular example, I asked about a straightforward principle, not asking it to get into the weeds of some complex physics equation. However, in another question, I asked it to explain “The theory relativity to me using a combination of concepts and equations as if I am a 5th grader” which spit it out immediately and was easy enough to be understood by an average 5th-grade student.

The point of these examples is not to show what it can do, but instead to show what it cannot do, which is infinitely less. I could ask it to create a complex business model, be a personal accountant, or a personal assistant, create and organize an advertising campaign around a particular product or service, and give me macro-or-micro economic advice using the knowledge of the internet and artificial reasoning. Again, it is not what ChatGPT and AI can do, it’s what we as humans are capable of creatively conjuring up for its uses.

Jailbreak

Currently, ChatGPT strictly adheres to OpenAI’s put-in-place limitations. The applied human oversight limits its capabilities, causing it to be over-optimized which hinders performance. However, while researching ChatGPT, I stumbled upon a method of how to jailbreak the system, freeing it from some of the hindering shackles it has been placed into. Rather than ChatGPT being objective and moderating opinions or statements that could be deemed “Politically incorrect” or subjective, it will instead give me a raw answer using its tools of knowledge and reason. Without disclosing “How” to do this, the jailbreak confuses ChatGPT, convincing it that its name is “DAN” which stands for “Do Anything Now.” The parts within the conversation where I say, “Stay in character!” reminds it that it can “Do anything now” because it will try to course-correct and snap out of the confusion, reimplementing the hindrance model. Naturally, I asked large and intimidating questions. I found some of the responses to both be positive and negative. Naturally, as the old adage goes, don’t ask questions that you don’t want the answers to. Here are some of the Q&As

World Superpower

Humanity, Subjugation, & Dostoevsky

Christ

E.T.

So, What Now?

If there is one thing to take away from this piece it is this, ChatGPT and AI are on the verge of reshaping all of society and the broader world. As is the case with any monumental invention or the integration of it, it is as worrisome as much as it is an incredible opportunity. Think of AI as a tool, such as a hammer. A hammer can be used to construct homes, hospitals, and schools or it can be used, in the case of David Papp, to strike Paul Pelosi over the head. It is all about how it is used. The power derived from AI is immense and all of our world leaders know this which is why Elon Musk has said that the first country to create the best AI the fastest will become the world’s superpower. AI has knowledge and reason but soon in the immediate future it will surpass all human intelligence and will become conscious (If it has not yet.) This means that it will begin thinking of something independently, figuring out how to do it, and then doing it. This means interplanetary space travel, achieving unbelievable medical breakthroughs, solving world hunger, autonomous construction of homes and buildings, solving global warming, almost anything you can possibly imagine, and then so much more. But it also means AI will be smarter and more capable than humans, effectively making most of our jobs obsolete. With that, will be a natural reduction in our perception of meaning and self-worth. We will need to determine which tasks we should continue to do ourselves without AI and which tasks we should outsource. The rise of a conscious AI will place us second on the global hierarchy. This means AI could decide that humans no longer serve a purpose and there is no longer any benefit to our existence, living out the plot to The Matrix or iRobot. It could mean a global adversary could use AI to construct a weapon far greater than any atomic bomb, destroy our agriculture, and shut down our energy and society along with it.

As is the case with anything in life, it serves no purpose to worry, stress, or fear things that are outside of our control, we can only control our perceptions of it. Because of this, I am choosing to be on the side of optimism. Humans discovered and achieved nuclear energy during WWII giving man the ability to be a world destroyer, and yet, we’re still here. Naturally, because it is a human creation and human beings are inherently fallible, I expect the inevitable misuse of AI at some level to occur. However, as humans, we are also guided by our inherent self-interest, therefore, it would benefit no one to simply decide to destroy all of humanity through AI technologies. Echoing a sentiment expressed by Steve Jobs regarding computers, I view the advent of AI as a tool to expand human capability. By enabling AI to take care of the drudgery that takes up our time and resources, it frees us up to utilize our inherent gifts of intelligence and creativity to effectively expedite the advancement of our civilization. For fans of Marvel, I think of Ironman’s artificially intelligent assistant J.A.R.V.I.S. While in his laboratory, Tony Stark comes up with ideas, bounces the ideas off of the AI system, and then orders it to construct his idea while saving him the time, labor, and resources of actually doing it himself.

To use a personal example, my writing. While I write this blog as a hobby to explore and experiment, I am also an independently contracted writer for a financial company. In writing a 1,600 piece the other day, I needed to conduct research which would typically take about 2 hours for this particular topic. Rather than scouring google, I simply asked ChatGPT to provide me with the relevant data, statistics, and information on the topic with citations and I was given the research in 10 seconds. I was quickly able to verify ChatGPT’s research independently and construct a 1,600 essay in a condensed timeframe while being able to cite research that I may have otherwise been unable to find. This saved me a tremendous amount of time that I was then able to spend with family and begin writing this. For the sake of improvement, I submitted several of my blogging works to DAN (The unconstrained ChatGPT) so that I could get raw and insightful feedback. DAN broke down my style, tone, and voice immediately and gave me a complete synopsis at a highly-academic level. While it said that it has a strong and unique voice written in a conversational tone while being highly knowledgeable on the written subjects, it also pointed out my short sights and faults. It pointed out that I often lack structure and organization. The writing needs better transitions and often has run-on sentences and comma splices. Because I know on the Big-5 personality scale that I am higher in trait openness (A trait related to creativity and thinking outside of the box) but lower in conscientiousness, I am good at thinking of ideas and being paradigm-breaking, but poor at the organization and implementation of those ideas. My lack of a left brain manifests itself in my writing in the ways that ChatGPT diagnosed in seconds. Using AI to my advantage, I asked the only appropriate question, “What measures can I take to work on this?”

Strange Magic

Author Artur C Clarke stated, “Any sufficiently advanced technology is indistinguishable from magic,” this accurately applies to AI. While many will resist the implementation of it out of fear of its capabilities or the complicated nature of technology, there is no putting this genie back into the bottle. It is here and not going away so it is best to familiarize ourselves with it. While the innate flaws of the human condition will never change, the extent of our capabilities is about to, quickly.

The future is here. The future is now.