The end of the AI hype

So is AI just one of many hypes? And will Gen-AI soon cease to be a major topic?

So the question arises: Is this all just hype? A term that promises a lot but delivers little? To assess this, it is worth taking a look at the so-called “Gartner Hype Cycle.” This model describes the typical phases of hype, from initial euphoria to realistic establishment in everyday life.

It starts with the “technological trigger”: initial successes attract attention, and the technology rapidly gains popularity. This is followed by a “peak of exaggerated expectations”: as awareness and attention surrounding the product increase, so do expectations, and a situation arises in which the media, investors, and companies often overload the technology with promises of salvation. However, these expectations can rarely be fulfilled. Almost inevitably, this leads to a “trough of disillusionment.” Projects fail, the benefits are not as great as expected, and the product is not the “miracle cure” it was thought to be. The result: users are disappointed because their expectations have not been met, and criticism eventually grows. Only then does the so-called “path of enlightenment” begin, on which expectations gradually decline and criticism of the product becomes increasingly widespread. Instead of exaggerated expectations, realistic and useful applications become established and the technology is used where it really fits. Finally, it reaches the “plateau of productivity” and becomes an established factor, albeit without the earlier hype, but with real added value for specific applications.

These trends are not mere theory. A look at past hypes shows that they repeat themselves regularly. A prominent example of this is IBM Watson. When the supercomputer impressed viewers on the quiz show Jeopardy in 2011, the media hype was enormous. Watson was celebrated as an all-rounder, especially in healthcare. But in practice, its use quickly failed. Clinics such as MD Anderson invested millions, but were never able to integrate Watson meaningfully into their processes. IBM eventually sold the Watson Health division for around a billion dollars, and the vision of a revolutionary medical AI largely vanished into thin air. The story was similar with big data: starting in 2012, it was celebrated in the media, research, and management consulting as a major future topic for bundling large amounts of data. But in most cases, it remained merely a matter of collecting data, and the actual use failed to materialize. Things moved even faster with NFTs: in 2021, the market reached a stock market value of around $17 billion, driven by speculation and media euphoria. Just one year later, 95% of this value has disappeared. Today, NFTs are still used in isolated cases, but little remains of the former hype.

But how can you tell whether a technology is just hype or whether the product is truly an innovation that will fundamentally change the market? Hype can often be recognized by typical signs: exaggerated success stories spread quickly in the press and investor circles. As a result, large providers suddenly come up with wish lists or there is massive VC and startup activity, but often without realistic business cases. The same patterns can often be seen: grand visions raise exaggerated expectations, but concrete use cases fail to materialize. The result: enthusiasm turns to disappointment and only a fraction of the original concepts survive. And the first signs of this pattern are already visible with GenAI.