Could the end of the AI hype cycle be in sight?

Like cryptocurrency and virtual reality before it, the new hype wave of generative AI may be losing some of its luster.

That’s according to a new set of tech predictions from analysts at research firm CCS Insight, who forecast a “cold shower” for the hype around AI in 2024 as “cost, risk, and complexity” start to bog down big dreams. The firm said risks around potential regulation and societal implications could also put a damper on early excitement.

“We think that the hype in 2023 has somewhat ignored some of the big obstacles that generative AI faces, and that ultimately, when people kind of wake up to the fact that there are huge challenges around this technology, it will slow progress in the short term,” CCS Insight Chief Analyst Ben Wood told Tech Brew.

Stuck between stations

While Wood predicts that deep-pocketed tech giants like Microsoft, Meta, and Google will continue to compete in the space, he expects developers and smaller organizations will soon wake up to the astronomical costs and compute power needed to make generative AI work.

“We’re seeing all of these exciting demos, like the eye candy of ‘I can get a unicorn flying over the White House with a dystopian theme behind it’—it’s fun,” he said. “But the costs of generating those kinds of images…are immense.”

There’s also the possibility of regulation adding more complexity to the AI equation. Wood said that’s likely to come from the European Union in the near term, as it tends to be a rulemaking “trailblazer” around tech-related issues.

“Companies will look at it and go, ‘Well, how far do we want to go down the route of delivering these capabilities if regulation is going to basically be behind us, and we may have to wind back some of the things we’ve done to be compliant?’” he said. “Because regulation is never going to keep up with the technology; the technology is going to be surging ahead. So it will be retrospective legislation.”

Add to that a flood of AI-generated spam content and images with iffy origins that might eventually give companies further hesitance about the safety of the tech, Wood said. “This is going to spark that debate about where we go with AI,” he said.

Big tech companies “can’t afford not to continue this arms race,” but they may face more questions from investors about ROI, Wood said.

“We have to move beyond some of these gimmicky implementations of generative AI, because they’re not going to be productive in any shape or form in terms of the resources they need, the cost—you can even talk about the environmental implications of cloud platforms churning out this stuff, which is a somewhat questionable value.”

Wood said CCS Insight remains “super positive” about the future value of generative AI, but it will take some time to cut the amount of resources needed to make it work.

“It feels like everyone’s getting a little bit carried away at the moment,” he said. “And it just needs to have a little bit of a reset.”