The Rise of Generative ChatGPT

The Rise of Generative ChatGPT

Across creative industries and disciplines, ChatGPT and other generative AI models and applications have sparked conversation after conversation around art, originality, and productivity. As this technology continues to dominate the public discourse and continues to occupy more and more space in creative and technical fields, we might stop and ask ourselves how this wave of automation came to be, and where exactly its roots can be found.

It is clear that ChatGPT and its contemporaries aren’t going anywhere, and learning where it comes from and where it’s going can be a big part of adapting to our new normal in the age of generative AI. Let’s take a quick look at the history of this technology, and its currently unfolding legacy.

A History of Generative AI

As it goes with any technology that threatens a paradigm shift, the roots and origins of ChatGPT stretch back much further than some people might think. Artificial Intelligence research as an academic topic dates back to the 50’s, coming first out of Dartmouth College. Really, though, the concept of autonomous content creation, especially of art and writing, can be traced at least as far back as Ancient Greek mythology. A western bent on history and philosophy often attributes a bit too much to Ancient Greece but, in this case, Hephaestus at least serves as proof that autonomous creation is not a new concept.

A big step towards our modern concept of AI came after the new millennium, as academics and other professionals began to develop more sophisticated statistical and generative models that would come to be the backbone of machine learning and language models. 2014 brought us the first deep neural networks with the ability to generate more complex data like images, and 2017 onward brought us various iterations of Generative Pre-trained Transformers—or GPT’s for short. 2021 showed us the birth of DALL-E, which was shortly followed by other AI art generators like Midjourney. GPT-4 represents the most advanced AI system to date, which is exciting for many and scary for many others.

The Explosion of AI Assistants

Automation is a powerful technology, no matter the scope. The industrial revolution marked an undeniable turning point in human history, bringing production into the age of machine automation or assistance, and blasting productivity into unimaginable levels. A world without this machine revolution is unfathomable to us now, but while this paradigm shift meant the amassing of capital for many powerful nations, it also shifted the consequences of untapped production to more vulnerable populations. Power, responsibility, and the usual implications.

Instead of factory productivity, the conversation now is around creative productivity. The introduction of AI assistants into popular discourse and technology brings into focus similar conversations. AI grants us the potential to boost creative productivity through the roof, but in a world that centers bottom lines over general well-being, automation is not always the most welcome prospect.

Still, AI assistants have shown up everywhere, from discourse to practice. AI Prompt Engineer is a new very well-compensated professional title that came with the widespread use of ChatGPT, and businesses left and right have been clamoring to see how ChatGPT and similar software can be put to work. And while concerns over exploitation and replacement are being validated by the SAG-AFTRA discourse, there is also an important conversation to have around the introduction of new tools, and workers’ ability to adapt to new technologies.

The Age of AI and ChatGPT

Unless some historical event veers us off our current path, generative AI and ChatGPT are here to stay, and will continue to embed themselves into every niche and discipline possible. There will be challenges and struggles to overcome, to be sure, but there will be and is already the necessity to learn how to navigate these technologies, and how to use them properly.

A common misconception today is the extent to which these tools can be trusted. OpenAI has been pretty open about how unreliable ChatGPT can be with certain prompts and in general, since it is first and foremost a model to approximate language and construct statistically likely and coherent responses. Ask ChatGPT to perform multiplication of large numbers and it will almost certainly be wrong. In this same vein, knowing what you can and can’t trust the model with is important, and knowing where it can be useful and where it can be harmful is just as important as knowing how to most effectively structure prompts.

That said, there are certainly effective and impressive applications of ChatGPT and other AI software that can be and have been implemented already. Understanding the limitations and true capacities of ChatGPT can be the difference between boosting productivity and producing misinformation but, with the right care and caution, ChatGPT can absolutely be used to enhance workflows for professionals and amateurs in many different fields and in many different ways. When it comes to generative ChatGPT, the question is certainly not whether it will bleed into all facets of our lives, but how willing we are to implement it safely and responsibly.