Is the Generative AI Bubble Going to Pop?

The world has been ignited with fervor for generative AI, a relatively new technology that allows sophisticated artificial intelligence (AI) to generate text and other media based on prompts from users. It’s one of the most impressive AI developments we’ve ever seen, and both businesses and individuals have been understandably clamoring to use it to its fullest potential.

But are we overhyping generative AI? And if so, is this excitement bubble going to pop in the near future?

The Basics of Generative AI

Generative AI takes many forms and is currently available in many different independent applications. But the basics are usually the same. Generative AI technology relies on large language models (LLM) to “understand” how language works and interpret information from a wide variety of sources.

Using this understanding, these tools are able to generate just about anything you request. This includes anything from a customer service call script to a screenplay for a movie about ninjas.

Mostly, people are excited about the following applications:

• Chatbots. Generative AI can be used to create and improve the capabilities of chatbots; so much so that you can have a full, human-like conversation with most generative AI tools. Arguably, these tools are better conversationalists than many humans. This makes them adept at automating things like customer service and help desk services.
• Content generation. People are also tapping generative AI for the sake of content generation. This includes creating blog posts from scratch or drafting email newsletters.
• Research and knowledge. Generative AI is also excellent for doing cursory research. It’s very good at summarizing large swaths of text and making complex topics easier for ordinary people to understand.

Of course, we’re only in the infancy of this technology’s development. We’ll likely see more creative, inventive, and game-changing applications in the future.

The Generative AI Bubble

If this technology is so impressive, how can we say that it’s in a bubble?

For context, economic bubbles tend to form when people overvalue assets. As an example, if housing prices in a given area rise in a way that’s disproportionate to the true value of those houses, it could result in overinflated valuations and consumers overpaying for houses. It’s usually only a matter of time before this bubble pops, causing prices to plummet back to normal.

If generative AI is in a bubble, it means people are overvaluing its functionality. That doesn’t mean that generative AI is useless, or that it’s unimpressive; it just means that there’s a significant gap between the real value of this technology and the perceived value of it.

With so many companies and individuals now heavily investing in this technology and hailing it as one of the greatest technological breakthroughs of our generation, it’s not hard to imagine how this gap could exist.

Why Generative AI Could Be Overhyped

These are just some of the concrete reasons why generative AI could be overhyped at this point:

• Lack of real-time information. Many generative AI tools don’t have access to real-time information. This makes them totally incapable of reporting or summarizing new events.
• Lack of opinions. Despite their fortitude at generating articulate, understandable text, AI isn’t conscious. It doesn’t think, feel, or form opinions. Many of us consume media because we’re interested in hearing the thoughts and opinions of the people creating it – not just a bland summary of a topic.
• Lack of creativity and novelty. While many people have tapped generative AI to complete creative works, like novels or scripts, this isn’t a good use of the tool. Generative AI lacks creativity and novelty, since it, by definition, can only mimic what it sees in the world around it. It might be able to replicate the stale jokes of repetitive sitcoms, but it’s not going to be able to come up with anything truly innovative.
• The human factor and user desires. Engaging with an AI might be suitable for certain applications, but most of us have a genuine desire to talk to and connect with other human beings. There’s a human factor that’s lacking in current generation AI, and that’s going to become apparent in applications like customer service in the near future.
• The need for guidance and supervision. Even the best proponents of generative AI admit that it’s borderline irresponsible to use text or media generated by AI without a thorough review. Additionally, effective media generation demands careful attention and prompt engineering. In other words, AI is nothing without human guidance and supervision.
• Long-term training problems. There are also long-term training problems that could emerge with generative AI systems, since eventually, these tools will run out of human sources to scrape and will begin training themselves on content generated by other AI systems. This AI feedback loop could lead to (and in some applications, already has led to) disastrous consequences.

What Would a Pop Look Like?

So what would it mean if the generative AI bubble actually pops?

It’s hard to say since this is such new territory. Companies or individuals that over-invest in this technology, or those that overly rely on it may be in for a rude awakening. It’s unlikely that people will lose money since these tools are relatively cheap compared to their apparent value, but misuse of these tools could easily lead to problems with authority and trust.

However, it’s worth mentioning that this technology is still developing and its long-term potential is much more interesting than its current capabilities. Even if generative AI is in a hype bubble, we are probably in for a wild ride into the future.