The journey of generative AI traces back to the 1950s when some academics introduced machine learning languages to ease their academic research process. The first mentionable generative writing AI project was ELIZA, developed in the 1960s with limited Natural Language Process (NLP) features.
The 2010s saw some development in AI writing tools, namely Jasper, Copy.ai, and Grammarly. These tools greatly improved content creation but were limited to paid users. But things changed overnight with the release of ChatGPT on 30 November 2022, a free generative AI tool that soon swept into various fields.
Since the release of ChatGPT, numerous AI applications for text, images, videos, and more have taken over the market like a storm. Almost every day nowadays, we hear the names of some new AI tools, which makes us feel overwhelmed.
As a result, a question often hits our mind – is generative AI a bubble like many other bubbles in the world? Can all these tools be profitable and survive in the market? Paralleling the past economic bubbles, I will discuss the impact of generative AI on the world economy.
Is Generative AI a Bubble? A Quick Hypothesis
In economics, bubbles are a situation when the price of any asset or market significantly goes up more than its natural value. It is caused by excessive speculation and investor enthusiasm. When the bubble bursts, the process falls and causes significant financial losses and economic disruptions.
It’s challenging to determine if generative AI is a bubble or not without witnessing the full development of the ongoing scenario. The suspicion arises because generative AI displays similarities to many past bubbles, like the dot-com bubble and the cryptocurrency boom.
However, waiting to see the end of this scenario isn’t a viable option, as many financial activities are already involved in its development. If it doesn’t seem to have enough financial prospects, this understanding could seriously impact decisions on whether we should invest time and money in generative AI tool development.
This is why I will cover this article based on several hypotheses. I will hypothetically consider generative AI as a bubble and present arguments for and against this hypothesis, referencing information available around to support my arguments and perspectives.
Before that, take a quick look at the history of generative AI development.
Evolution of Generative AI – Brief History
The development of generative AI is an extraordinary story of the human ability to harness the imagination with technology. From its conceptual inception to recent effects, generative AI has evolved into a powerful force to change the world.
The Early Days of Generative AI
As said above, the research for developing generative AI actually began in the late 1950s, with machine learning (an emerging theory that allowed algorithmic creation of new data).
Perhaps the first was the Markov Chain, a statistical model that produces rows and columns of data given probabilities in input. But in those early days, there was limited computing and data processing power.
Until the 1990s-2000s, machine learning couldn’t come to its own as the improvements for hardware and digital data weren’t widespread.
Explore this post – is WordPress a SaaS? The debate is explained.
The Turning Point: Neural Networks and Generative Models
The idea of neural networks gave the development of modern generative AI a new dimension. In this system, computer simulations drew inspiration from the human brain. These ‘connected neurons’ helped machines sort and learn from huge amounts of data.
In 2014, Goodfellow and colleagues invented the General Adversarial Networks (GANs) that remarkably revolutionized the field. It functions by contrasting two neural networks: the first creates data, and the second checks whether it’s true. This process opened up the possibility of producing extremely accurate and imaginative products.
Around the same time, other generative models like Variational Autoencoders (VAEs) and Recurrent Neural Networks (RNNs) came into play, opening the door for the incredible expansion of Generative AI.
Key Milestones in Generative AI
In the past 10 years, Generative AI has achieved some incredible milestones, each of which is an advance in capacity and use:
- WaveNet (2016): Designed by DeepMind, WaveNet changed the face of text-to-speech by producing human-like speech at an unprecedented scale.
- Progressive GANs (2017): NVIDIA’s technology revolutionized the way visual content is created with high-resolution, photo-realistic images.
- GPT-2 and GPT-3 (2019, 2020): OpenAI’s Generative Pre-trained Transformer models became standard for A.I. text generation, providing consistent and contextualized output for chatbots, writing programs, etc.
- DALL-E (2022): A mix of visual intelligence and creativity, DALL-E brought with it the power to create detailed pictures out of natural language commands.
- ChatGPT (2022): OpenAI’s conversational AI service, which became extremely popular, achieving a million users in 5 days.
- GPT-4 (2023): As the most sophisticated AI ever developed, GPT-4 was capable of higher-level reasoning and accuracy, which confirmed the role of Generative AI in many fields.
Most Popular Generative AI Tools
Below, I have listed some of the best and most popular generative AI tools.
Writing Tools
ChatGPT, Jasper, Copy.ai, Writesonic, Rytr, Grammarly, and You.com.
Designing
DALL-E, Canva, Midjourney, DeepArt, Leonardo, Artbreeder, Runway ML, and Designify.
Coding
GitHub Copilot, Tabnine, Kite, Codex, Replit, Codeium, and Sourcegraph.
Data Analysis
DataRobot, Tableau, Alteryx, RapidMiner, BigML, TIBCO Spotfire, and Looker.
Automation
Zapier, UiPath, Automation Anywhere, Microsoft Power Automate, Integromat, Pipedream, and Katalon.
A Short Picture of the Generative AI Economy
Generative AI is going to re-engineer the world economy in ways that have profound implications for industries, regions, and economies. This generative AI market is forecasted to grow unprecedentedly, from $23.1 billion in 2024 to $440 billion by 2032, speaking to its revolutionary economic potential.
Significant Contribution to Global GDP
Generative AI has already been adopted in all industries, creating a comprehensive economic value. This can exceed the economic growth of some major economies of the world and increase GDP by $7-$10 trillion, covering as much as 10% of the global economy. It can also save 300 billion hours a year.
Generative AI technology is expected to expand China’s GDP by 26% (or $10.7 trillion) by 2030 and North America’s by 14.5% ($3.7 trillion). These gains reflect AI’s contribution to regional economic leadership and innovation-driven development.
Transformation Across Sectors
Service industries such as healthcare, education, software development, and administration benefit enormously from generative AI technologies. Many companies are adopting AI in their core plans. In 2024 alone, businesses around the world invested around $4.6 billion into generative AI applications.
Societal Integration
The use of generational AI extends to the people: 4 out of 10 US adults (18-64) have adopted using it in their workplace. This spread is the evidence of its utility and availability, adding to its economic power.
Explore the post of best web cloud hosting providers.
Is Generative AI a Bubble? Hypothesises Explained
Now, I will explain some popular and considerable hypotheses as to why generative AI is suspected to be a potential bubble.
Hypothesis 01: Rapid Investment and Valuation Surge
Generative AI has raised billions in venture capital in recent years, with some startups that value only millions of dollars. This vast investment happened, although most firms could not establish sustainable business models or revenues. This indicates investors are taking a gamble rather than checking how financially healthy or market-ready these firms are. For example:
OpenAI
In early 2023, OpenAI received a massive investment from Microsoft, amounting to $10 billion. Microsoft created a partnership and integrated OpenAI into their products.
This investment increased the valuation of OpenAI to $29 billion, although its revenue was primarily timed to its API services and product like ChatGPT, which were only in the early stages of monetization.
Anthropic
It’s an AI safety and research company that raised $580 million in a Series B funding round in 2023 from companies like Sam Bankman-Fried’s trading firm and other prominent investors.
This funding increased its valuation to nearly $4 billion despite its limited commercial products at that time.
Stability AI
Stability AI is well-known for developing the Stable Diffusion model. It raised a $101 million investment in 2022 led by investors like Coatue Management and others.
This funding soared its value to $1 billion, although it kept relying on open-source and wasn’t making any considerable revenue then.
The same goes for AI products like Jasper, Copy.ai, Runway ML, and LLaMA.
Hypothesis 02: Overhyped Expectations
Generative AI’s technical potential and impact are frequently overblown to create unrealistic market hopes. Media reports and promotional statements often portray generative AI as a disruptive technology that will completely transform industries.
The fact remains that generative AI can deliver impressive content, design, and coding results, but the technology is not entirely perfect in all respects, including bias, quality control, and ethics. This mismatch between expectations and reality may ultimately frustrate investors and users.
For example:
When generative AI systems such as GPT-3 were first introduced to the market, there was hype that they had the ability to do everything from content creation to customer support without needing any human help.
However, as businesses started to use these tools, they found that these tools have many limitations and often provide inaccurate information that can severely damage the quality of research-based works. As a result, organizations have started hiring human workers again in the workplace.
Hypothesis 03: Limited User Adoption
Even a few months ago, there was a serious buzz around companies, organizations, and professionals adopting the premium versions of the generative AI tools. But since their limitations started being evident, most of them are seen not to renew their premium subscriptions.
They can easily carry out their regular tasks with the free versions. Only those who operate in sophisticated workspaces have been seen to use the premium versions. They are a very small percentage of the global population. This has seriously affected the revenue stream of these generative AI tools, making their financial prospect more volatile.
Here’s a guide on how to find the best and cheapest web hosting plans.
Hypothesis 04: Market Saturation
Can you tell me the total number of AI tools that are now out there online? Yes, it’s really difficult to say. Because almost every day or week, some new tools for each industry are coming out, making the entire market highly saturated. According to ArtSmart.ai, there are now 2000+ generative AI tools online.
As a result, it has become difficult for one AI tool to differentiate itself from the other and propose USPs since most of them offer the same set of features and functionalities. If people want to adopt any AI tool, they will love to select the ones that are most familiar and globally established.
This saturated market can be less profitable for investors as new startups may find it difficult to gain an acceptable market share.
Hypothesis 05: Technological Limitations
Current generative AI technologies might not be sophisticated enough to live up to their exalted claims. Generative AI systems often draw on huge amounts of data and complex algorithms, which may lead to biases and moral problems, which we have already explained above.
Also, the output quality is unpredictable, which makes users unsatisfied. Without overcoming these technological hurdles, interest and investment will decline.
Hypothesis 06: Economic Factors
The general economic picture, like a recession or a change in investor mood, could lead to a decline in investment in generative AI projects. The impact of economic downturns may cause a reduction in venture capital investments as investors lose their risk appetite.
Such a recession might cause most generative AI startups to be unable to raise the funding needed for expansion, let alone survival. Historically, speculative investments have tended to bleed in economic downturns.
Hypothesis 07: Comparison to Previous Tech Bubbles
We can find many similarities in the ongoing generative with the dot-com bubble of the late 1990s and early 2000s. During the dot-com bubble, many startups with ill-established businesses and overpriced valuations fell as the market bounced back to its original state.
Likewise, since generative AIs are experiencing a large rise in investment with no strong fundamentals, they could suffer a massive collapse in valuation once the hype wears off.
Closing Up!
Hope you will agree that all the hypotheses I have presented above carry enough significance to consider why generative AI might be a potential bubble. But don’t take my claim/hypotheses for granted. I am trying to warn you based on my knowledge of previous global economic bubbles.
If it doesn’t appear as a bubble in the next five years, then it’s really cool. Our world will be highly benefitted. But if it turns out to be a bubble, million to billion-dollar investments would be a waste. So, it’s better if you make your investments with enough caution.
The sky is cloudy, but that doesn’t mean there will be rain today. Yet, you must carry an umbrella for an extra layer of caution. The same is the gist of my today’s post. If you have any comments, please mention them below.
