It’s been two years since the term “generative AI” started flooding my email inbox. It wasn’t a new term (it appeared in one of Gartner’s famous Hype Cycle reports back in 2020), but as the summer of 2022 drew to a close, the incoming messages and pitches I was receiving were a clear sign that hype was quickly building around AI-powered tools that could generate content – like text, images, and computer code. And when OpenAI launched ChatGPT in November 2022, generative AI catapulted into mainstream culture and has been on a roll ever since.
But something has changed in this cheerful story in the last few weeks.
Goldman Sachs called generative AI “overrated” and “insanely expensive.” Venture capital firm Sequoia Capital said “the AI bubble is reaching a tipping point.” A barrage of media headlines like “The AI hype machine is running on reserve” eagerly pour cold water on the whole issue.
Why? Well, generative AI chatbots struggle to answer simple questions or feign false information. The most sophisticated generative AI models are constantly hungry for data and processing power. Generative AI startups that generate little to no revenue must constantly seek massive funding rounds to stay afloat. Fortune 500 companies cannot bring generative AI use cases into production because of concerns about accuracy, liability, and safety.
And with the S&P 500 experiencing its biggest sell-off in two years on Monday, the feeling that the generative AI bubble is beginning to burst is growing.
According to Gartner’s Hype Cycle, generative AI has passed the “peak of inflated expectations” and is now heading toward an impending “trough of disappointment.” If that’s true, what comes next will be painful and disruptive. Investment money could dry up. Startups could fail. There could be layoffs.
For many of the startup employees, founders and investors who have invested and taken the work and risks necessary to get the generative AI sector off the ground, the sting of the market correction will be unfair and brutal. But knocking generative AI off its high pedestal is also necessary for the long-term sustainability of the AI landscape, Kjell Carlsson, a former analyst at Forrester Research who is now head of AI strategy at enterprise data platform Domino Data Lab, told me.
“I’m pretty confident that people will realize that Gen AI is not just AI,” he said, pointing to the wide variety of other artificial intelligence technologies, including predictive AI and machine learning, that were already delivering real return on investment before the advent of generative AI. “Gen AI is a set of technologies that are part of this broad toolkit of different technologies that require work,” he explained. “There’s no magic button, it’s about leveraging technologies for the right use cases.”
Don’t be afraid of the trough
To be clear, generative AI is not going away. These models and tools, from ChatGPT and Microsoft Copilot to Google’s Gemini, Anthropic’s Claude and Meta’s Llama, have already become part of our lives – for productivity, efficiency or just for fun. Just as we have become accustomed to getting all the information we need in seconds via a Google search, so too will the ability to get easy-to-read summaries of work meetings, compose memos to colleagues and create images and presentations by speaking just a few words.
But let’s also be realistic: the huge investment in generative AI, estimated at a trillion dollars, has yet to pay off. Much of it may not be as absurd as, say, the dot-com bubble of UrbanFetch and Pets.com (I still remember the ice cream deliveries and doll giveaways), but it’s hard to argue against the idea that generative AI is getting the reality check it deserves.
“The irony is that I think I was the first industry analyst to jump on the Gen AI bandwagon,” Carlsson said. “Although it was a success by anyone’s standards, the expectations of how quickly it would impact the bottom line of large organizations did not match the reality.”
This is where the so-called valley of disappointment becomes an important phase for any technological development, said Chris Howard, global research director at Gartner, in a recent video. The premise is simple: After an initial burst of excitement and enthusiasm among early adopters, the new technology makes its way into the hands of mainstream users, who find that it doesn’t live up to their inflated expectations. A retreat follows, during which the technology is refined and expectations are reset.
“It’s not this dark, dangerous place,” Howard explained in the video. “It’s where we figure out how to make something work – or not.”
For generative AI, the low point will be a period marked by small, incremental advances in applications that bring real benefits to companies and users. Less well-known, however, will be declarations by OpenAI CEO Sam Altman that he has created artificial general intelligence (AGI), “the most powerful technology ever invented by humanity” – even if that may make for less sexy headlines.
Even Dan Ives, a Wall Street tech analyst at Wedbush who remains bullish on AI stocks, said this is a key period for tech companies to walk the walk, not just talk the talk, when it comes to generative AI. They need to “show the use cases and monetization to justify the AI revolution,” he wrote to me in a text.
Ives said he believes Microsoft, AMD, Nvidia, Palantir and Oracle have shown they can deliver real value. Still, the sector as a whole still has a lot to prove, with so many generative AI startups riding on multibillion-dollar valuations.
There are no guarantees, but there is a long history of mature AI technologies that have contributed to other, newer AI disciplines, such as computer vision, which has become a central part of today’s multimodal generative AI (AI that can generate not only text but also images and videos, for example).
Perhaps generative AI, driven by other, newer technologies such as agent-based AI (AI systems that pursue complex goals and workflows like autonomous agents), can still reach its full potential.
Now it may be time to start the real, hard work on generative AI. “I think this is going to be a false AI winter,” said Steve Jones, executive VP at technology consulting firm Capgemini, in a LinkedIn post. “A winter where hopefully the hype will be over and we can focus on getting our work done.”
CEO Daily provides the most important context to the news that leaders across the business world need to know. Every weekday morning, more than 125,000 readers trust CEO Daily for insights into the C-suite and its surroundings. Subscribe now.
