One Year Of ChatGPT: Five domains disrupted by generative AI

The first change triggered by the mass rollout of ChatGPT was automation, or rather, our understanding of who it will affect.

ChatGPT is not just a tool of the present. It is a small glimpse into the future of human-AI interaction. Since its official launch one-year prior, ChatGPT has transcended the realm of tech to become a cultural phenomenon.

It was the first true AI system that showed us what LLMs (Large Language Models) are truly capable of, and it made us collectively make a mental leap into the future. The excitement, consternation, and debate that it induced has permeated nearly all sectors and stakeholders.

And today, as the first anniversary of this seminal technology’s release rolls in, it would be useful to reflect on all that has changed in the past year.

Automation

In no particular order, the first change triggered by the mass rollout of ChatGPT was automation, or rather, our understanding of who it will affect. Pre-2022, it was widely believed that the first victims of automation would be blue collar roles, as their jobs demanded lower skill and they were repetitive in nature.

However, OpenAI completely flipped the script on this and showed that white collar, and especially creative roles, were far more at risk of automation than anyone had imagined. It was a rude awakening for many of us, initiating new discussions on automation, productivity, and universal basic income, while also forcing the world’s media to be more cognisant of new developments in AI.

Education

The next sector that has been forever altered is academia. Cramming information to reproduce in a test, writing essays, and selecting the right options in multiple-choice questions, has been the usual pattern of education and testing for centuries.

But now, ChatGPT has changed all that. It can write essays better than most of its users, presents coherent information from a diverse array of sources, and performs exceptionally well on various standardised tests. This has upended the paradigm of education that has been with us for generations, raising various philosophical and tangible concerns along the way.

How to attune the current system of education to adjust to this new reality? What is the purpose of training students in a field where most will never surpass the capabilities of AI? And what is the fair amount to charge for a multi-decade long educational process in this new era of AI? All these will need answers soon.

Geopolitics

A third and more contemporary arena that ChatGPT changed in this year is geopolitics. OpenAI’s product showed governments around the world that AI is going to be one of the pivotal technologies of this century, and this has already led to strategy formulation and multilateral discussion around AI.

Take the recognition of AI as a field of competition between US and China, the multilateral conference on AI safety that took place recently in the UK, or the passing of the EU’s AI Act. ChatGPT has stirred a discussion on AI in the international relations space that is expected to only intensify further.

Top AI engineers of today will be seen as strategic assets in the same way that nuclear engineers were during the Cold War.

Semiconductors

Another sector that ChatGPT impacted, albeit indirectly, is semiconductors. This is due to the computational demands of training and deploying such large and powerful models.

The semiconductor pipeline is one of the most concentrated supply chains in the world with Nvidia, TSMC, ASML, Samsung, and a few others cornering a lion’s share of the supply. While OpenAI’s models were trained on Microsoft’s cloud and relied on Nvidia’s A100 chips primarily, the success of ChatGPT has made many more firms interested in developing their own models and chips, initiating change in this hyper-concentrated industry.

Currently, Amazon is making their own chips, AMD is getting into the game, Qualcomm is upgrading its chips for AI tasks, Google has built the Tensor chip, and so on. Obviously, some of these developments predate ChatGPT. Yet, it cannot be argued that the recent thrust that they are seeing is attributable to the buzz created by ChatGPT.

Software & KaaS

Embedded AI in mainstream applications is also something that ChatGPT could be thanked for. Whether it is directly, such as Microsoft Windows or their Office Suite getting upgraded through a “Copilot” which is powered by ChatGPT, or induced, such as Adobe integrating generative AI into their creative cloud suite, the expectations from apps has been forever altered.

Now more and more applications have embraced this trend, and it appears that generative AI could soon become as ubiquitous as the internet. With hundreds of millions of users garnered in less than a year, one shouldn’t bet against it.

Another dramatic shift that ChatGPT induced was introducing the concept of “Knowledge as a service”. The models that power ChatGPT and a host of other generative AI tools, have an underlying neural network that stores information. This knowledge can then be accessed at will to gain insights or drive solutions. ChatGPT’s ability to provide accurate and curated information on demand has led to the creation of this space overnight. Now, with a host of firms demanding such capabilities in house, and recent upgrades allowing the creation of custom ChatGPT’s, the domain of “KaaS” is only set to expand.

Tip Of The Iceberg

All the above points capture only a small subset of changes that ChatGPT induced, and only serve to illustrate how the world around us has evolved with the release of this tool.

Numerous other items could also be added to this list such as the new challenge of skilling workers for the age of AI, copyright issues related to training data, the issue of deepfakes and disinformation, generative AI serving cybercriminals, and much more.

In many ways, it feels like the early days of social media. We know that generative AI will change various aspects of our lives, and the world will never be the same. But how many spheres will be impacted, and to what degree will things change, we can barely even fathom.

Srimant Mishra is a computer engineer and lawyer with a deep interest in the field of Artificial Intelligence. Views are personal, and do not represent the stand of this publication.