Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More

---

Is it possible that the generative AI revolution will never mature beyond its current state? This question arises from deep learning skeptic Gary Marcus, who recently suggested that the generative AI “bubble has begun to burst” in his blog post. Generative AI refers to systems that create new content based on patterns learned from vast amounts of existing data. Recent news stories and analyst reports have raised questions about the immediate utility and economic value of generative AI, particularly for bots based on large language models (LLMs).

Historically, skepticism about new technologies is not uncommon. For instance, Newsweek published an article in 1995 claiming the Internet would fail, arguing that it was overhyped and impractical. Today, as we navigate a world transformed by the Internet, it’s worth considering whether current skepticism about generative AI might be equally shortsighted. Are we underestimating AI’s long-term potential while focusing on its short-term challenges?

Goldman Sachs recently questioned the value of generative AI in a report titled: “Gen AI: Too much spend, too little benefit?” Additionally, a survey from Upwork revealed that nearly half (47%) of employees using AI are unsure how to achieve the productivity gains expected by their employers, with 77% stating these tools have decreased their productivity and added to their workload.

A year ago, Gartner listed generative AI at the “peak of inflated expectations,” but more recently indicated that the technology is slipping into the “trough of disillusionment.” This phase is characterized by waning interest as experiments and implementations fail to deliver results.

While Gartner’s assessment points to disappointment with early generative AI, this cyclical pattern of technology adoption is not new. The buildup of expectations, often referred to as hype, is a natural human behavior. We are drawn to new technologies and their potential, yet early narratives are often misleading. Realizing that potential requires hard work and rarely goes as smoothly as anticipated.

Analyst Benedict Evans recently discussed the complexities of AI adoption, stating that it takes longer than expected to translate utopian dreams into reality. This overestimation of new systems is at the heart of technology bubbles.

Roy Amara, a Stanford University computer scientist, famously noted in 1973 that we tend to overestimate the impact of new technology in the short run while underestimating it in the long run. This observation, known as “Amara’s Law,” remains relevant today.

The maturation of new technologies often takes time. Ken Olsen, CEO of Digital Equipment Corporation, once claimed, “There is no reason anyone would want a computer in their home,” years before personal computing became ubiquitous. It simply required time to evolve.

The likely progression of AI technology

Given this historical context, it’s fascinating to consider how AI might evolve. A 2018 study by PwC outlined three overlapping cycles of automation driven by AI, extending into the 2030s: the algorithm wave, the augmentation wave, and the autonomy wave.

This projection aligns with current discussions on AI’s role in augmenting human abilities. For instance, IBM’s first Principle for Trust and Transparency states that AI’s purpose is to augment human intelligence. An HBR article explores the relationship between humans and AI, emphasizing the potential for AI to enhance creativity.

There are already numerous examples of AI augmenting human capabilities. In healthcare, AI-powered diagnostic tools improve disease detection accuracy, while in finance, AI algorithms enhance fraud detection and risk management. Customer service also benefits from AI through sophisticated chatbots that provide 24/7 assistance.

However, augmentation does not equate to full automation of human tasks, nor will it eliminate many jobs. The current state of AI resembles other computer-enabled tools, such as word processing and spreadsheets, which enhance productivity without fundamentally changing the world.

Short of expectations

The hype surrounding generative AI often leads to the expectation that it is revolutionary or will be soon. The gap between these expectations and current realities fosters disillusionment and fears of an AI bubble bursting. What’s often overlooked is the realistic timeframe for development. Evans recounts a story about venture capitalist Marc Andreessen, who believed that every failed idea from the Dotcom bubble would eventually succeed — it just took time.

AI development will continue to progress, albeit unevenly across industries and professions. There will be ongoing examples of impressive gains alongside stories of AI technology falling short. This uneven trajectory characterizes its awkward adolescent phase.

The AI revolution is coming

Generative AI will indeed prove revolutionary, although perhaps not as soon as some experts predict. The most significant effects of AI are likely to be felt in ten years, coinciding with PwC’s anticipated autonomy wave, where AI will analyze data, make decisions, and take actions with minimal human input.

As we approach this autonomy wave, we may see AI applications becoming mainstream, such as in precision medicine and humanoid robots. Today, AI is already augmenting human capabilities in meaningful ways. The AI revolution is unfolding before our eyes, albeit more gradually than some anticipated. Perceived slow progress could lead to pessimism about AI’s future, but in the long term, in line with Amara’s law, AI will mature and fulfill its revolutionary potential.

Gary Grossman is EVP of technology practice at Edelman.