The Chatter Behind ChatGPT: How AI Is Leading in Industry Innovation

You’ve probably read it all over news stories, company chats, and LinkedIn feeds. Your family member is curious about it, or your boss wants to implement it. It’s ChatGPT, the latest chatbot fueled by artificial intelligence and launched by Silicon Valley startup OpenAI.

From a technological perspective, writing devices like ChatGPT are an incredible leap and promising gateway for the future of AI, and specifically, large language models. However, the question on everyone’s mind is: what does it mean for us as humans?

Now, joined by other AI writing tools like Google’s Bard, Jasper.ai, and Canva Magic Write, ChatGPT is poised to change the way we view and produce content. Let’s take a closer look at the latest disruptive technology claiming our thoughts and conversations.

What Is a Disruptor?

Many news outlets and industry leaders have called ChatGPT and its parent company OpenAI a “disruptor.” Disruptors are companies that have the potential to change or topple existing companies and industries. Whether they hold innovative technologies or create more efficient operations, they far surpass the old ways of doing business. While disruptors can be found across many areas of the economy, it’s helpful to think of five key areas of disruption: automation, communications, finance, medicine, and technology. Many of the current disruptors are building out capabilities for the metaverse, blockchain technology, and autonomous driving, to name a few.¹

Prominent models for disruptors are Amazon, Netflix, Airbnb, and Uber. Netflix, for example, made at-home viewing easier than ever and upended video and DVD rental businesses around the world. Now, it has opened an entirely different industry of streaming services and connected TV, with a new crop of competitors, that in turn has shaped the way we make TV and movies (e.g., limited series and live comedy specials). OpenAI and its frontline product, ChatGPT, are the latest to join this elite group.

There are benefits and drawbacks to these influential innovations. On one hand, they can eliminate inefficiencies, reduce costs, expand access, prioritize convenience, improve safety, or change people’s thinking. For ChatGPT specifically, OpenAI states that their mission is to “ensure that artificial general intelligence…benefits all of humanity.”²

However, as with many rising stars, it means that others must fall. After video killed radio, then Netflix killed BlockBuster. As the term implies, entire companies and industries are “disrupted,” which can lead to financial losses, job displacement, and uncertainty. Regulating a new product or technology about which very little is known is difficult—its dangers might be discovered far too late, when it’s already been adopted by the masses. And, as always, the constant marathon toward innovation begs the question: is new always better?

The History of ChatGPT

OpenAI, an artificial intelligence research non-profit company, was founded in 2015 by Sam Altman, Elon Musk, and other Silicon Valley investors. Together, they hoped to “develop safe and open AI tools to empower (rather than eradicate) people.”^3,4 In 2016, they started by focusing on AI for video games and other applications, until they introduced the concept of a Generative Pre-Trained Transformer (GPT) in 2018. GPTs are neural networks that are inspired by the structure and function of the human brain, then are trained on large amounts of human-generated text to act like one.⁴

In 2019, OpenAI released a chatbot that could engage with users in casual conversation. Unfortunately, it was so good at writing fake news stories that they decided it was better to not release it. But, later that year, the company released an updated version of the tool, known as GPT-2, and announced a $1 billion partnership with Microsoft to invest in and expand its capabilities.³ ChatGPT-3 came shortly after in 2020, followed by its iterations ChatGPT-3.5 in November of 2022 and ChatGPT-4 (with the widely known ChatGPT) product in March of 2023.⁵ Within five days, ChatGPT had one million users and became the fastest-growing Internet service ever.⁶

In 2015, OpenAI changed its status to a "capped profit" company, meaning that it cuts returns from investments past a certain point. To reflect this shift, they changed their name to OpenAI LP under a new charter: “investors and employees can get a capped return if we succeed at our mission, which allows us to raise investment capital and attract employees with startup-like equity. But any returns beyond that amount—and if we are successful, we expect to generate orders of magnitude more value than we’d owe to people who invest in or work at OpenAI LP—are owned by the original OpenAI Nonprofit entity.”⁷

The Power of AI Chat

ChatGPT is an AI-powered chatbot that can simulate human conversation, write code, and produce hundreds of words in seconds. Once you write a question or prompt to the bot, it will answer back in seconds. According to OpenAI’s website, it can “generate, edit, and iterate with users on creative and technical writing tasks; accept images as inputs and generate captions, classifications, and analyses; and is capable of handling over 25,000 words of text, allowing for use cases like long form content creation, extended conversations, and document search and analysis.”⁸

With ChatGPT’s help, you can craft an important email to your boss, code a new website, or find a recipe that uses the ingredients in your pantry. Unlike a search engine, ChatGPT can’t scour the Internet to find exact information that you’re looking for. However, it will engage in conversation using its training data to give you a response—perhaps with a few new ideas along the way.

OpenAI has also designed and created DALL-E, a tool for creating images based on a prompt, and Whisper, which can transcribe speech into text and translate many languages into English.⁸

How It’s Informed

To help provide the exact types of answers users are looking for, ChatGPT relies on large language models (LLMs) and reinforcement learning from human feedback (RLHF), as well as information found in its training data up to 2021.^5,9 Before it was launched, every employee at OpenAI tested its flaws to try to break the model. Upon its release in March of 2023, OpenAI made the tool free and open to use so that it could solicit feedback from millions of users at once and continue to improve.⁵

What Makes ChatGPT So Special?

Efficiency, speed, and ease of use are three of the biggest benefits for ChatGPT users. Its apparent universal appeal and current free, global access are also noteworthy, but are still in debate and may change in the very near future because of its infancy.

“Advanced language models like ChatGPT that use natural language processing and GPT-3 technology can generate large amounts of high-quality content at a fraction of the cost and time it takes for humans,” says Lanie Shalek, director of growth marketing at Jobi Capital.¹⁰

Although ChatGPT is not a form of Artificial General Intelligence (AGI is a machine that can perform any task that a human can), it can create content that closely resembles, and at times fully mimics, human output.¹¹ This eliminates mundane or redundant tasks so that people can focus on larger initiatives.

“One of the most audacious things about the LLM revolution is that almost no one’s work is spared a potential productivity boost,” says Cassie Kozyrkov, Chief Decision Scientists at Google.¹²

ChatGPT also is impressively straightforward and simple to use. Users simply type a prompt into a text box and wait for the chatbot to respond. If the answer is insufficient, you respond again and clarify what you’d like to see. According to John Schulman, a co-founder of OpenAI, the latest version of ChatGPT readily infers intent, and users can get to what they want by going back and forth with the bot.⁵

Another major benefit is its vast variety of use cases. As mentioned before, users can lean on this tool to help them write, brainstorm, compose, analyze, summarize, learn, design, and more. It can help solve new and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more, without needing significant resources or time.¹¹

Challenges & Obstacles

Although many people are concerned about the ways in which AI can replace us, especially at work, it’s important to focus on the ways in which it doesn’t, and likely won’t ever, reach full human potential.

First, ChatGPT is not a human and will never be one. Its “uncanny valley,” or feeling of strangeness/fear around technology that is almost humanlike but not entirely, can’t be breached without significant evolutions.¹⁰ The chatbot can regurgitate content on the human experience, but it can’t sympathize, replicate, or imagine it.

“I can tell when text is AI-generated right now, because there’s a quality to it that doesn’t quite fit; it doesn’t quite make sense,” said Ajay Goel, founder of email marketing platform GMass.¹⁰

This raises additional issues of copyright infringement, plagiarism, and outsourcing work. What should AI writing be used for? If students use the tool to write a paper, or job applicants plug in a prompt to return to a company, can they still take credit for their work? Who needs to disclose that they’re using ChatGPT and when?

In terms of personal and professional development, ChatGPT also cannot replace thinking. People who use an AI writing tool “have not done the hard work of taking inchoate fragments and, through the cognitively complex process of finding words, crafting thoughts of their own,” says Johann N. Neem, professor at Western Washington University.¹³ Removing the thought process removes opportunities for mind wandering, daydreaming, and independent research, which often lead to epiphanies and breakthroughs. Thus, it also cannot generate completely original ideas or content, which are the foundation of remarkable work and art.

Lastly, and perhaps the most pressing of all, is ChatGPT’s inability to discern between fact and fiction. OpenAI admits that accuracy and misinformation is a priority, and that they plan to actively test ways to break and improve their content moderation policy. Although the current content policy is set to flag and remove unsafe or illegal information, some users have been able to elicit questionable responses or written work that is proven to be false.⁵

“From my perspective, ChatGPT fails a lot—there’s so much stuff to do. It doesn’t feel like we’ve solved these problems,” says John Schulman. “We all have to be very clear to ourselves—and to others—about the limitations of the technology.”⁵

The Future of AI Chat

With all of this attention and excitement, where will ChatGPT go next? In the short term, OpenAI is hoping to refine its RLHF approach, expand its use cases, and minimize jailbreaks, according to its executives. After signing deals with Microsoft (worth $10 billion), Buzzfeed, Bain, and other high-profile organizations, they also want to see how the chatbot can perform in different environments.⁵

ChatGPT’s future will also be influenced by its competitors, the largest of which was announced mere days after GPT-4: Google’s AI chatbot Bard. Meta, having already tried and failed at launching a chatbot, likely isn’t far behind.⁶ With more products entering the space, there is sure to be more regulation and oversight from industry and government agencies as well.

Our Perspective at New York Tech

As a technology-focused university deeply infested in advancement and opportunities, New York Institute of Technology recognizes the magnitude of AI like this and is working to educate our faculty and students on how to incorporate, detect, and moderate it. Since the release of ChatGPT and other similar tools, we’ve been conducting workshops on how to bring AI generators into the classroom to engage students in meaningful work and challenge their critical thinking.

In these workshops, we discussed:

• Where can ChatGPT facilitate a better outcome?
• Are there efficiencies that can be gained in grading?
• Are new rubrics and assignment descriptions needed?

As always, New York Tech learns from the latest technologies and incorporates them into the curriculum. We already bring many new tools, programs, and processes into our Online Data Science, M.S. program, and we will continue to do so as AI evolves.