The generative AI sector will take a “cold shower” next year as the cost of operating these services begins to weigh on businesses, an analyst firm has warned.
Generative AI has undoubtedly been the most important technology of 2023, but CCS Insight predicts that this year’s hype will be replaced by the harsh reality of cost and risk calculations in 2024.
“Already this year there has been a sense of transition,” said Leo Gebbie, principal analyst at CCS Insight. “At the start of the year, the talk around generative AI was all about the exciting promises of the technology and what it could do, what it could deliver, the incredible things it could do to We.”
However, the expenses of running millions and millions of queries are starting to add up for the big AI players, and it’s likely that these costs will be passed on to consumers and businesses at some point. “The cost of running generative AI-style queries in the cloud is much higher than running traditional search queries in the cloud,” Gebbie said.
“We think there will be a little cold shower next year,” he added.
Deficit AI
A report in The Wall Street Journal published earlier this week claimed that Microsoft was losing an average of $20 per user on GitHub Copilot, the company’s code-writing AI, earlier this year. That could explain why the company has set a relatively high price of $30 per user for the Microsoft 365 Copilot AI it will launch next month, which brings generative AI capabilities to apps like Word, Excel and PowerPoint.
“Microsoft is a really interesting example, where they very directly produce Copilot and charge an additional amount for it, but it’s not an insignificant additional amount,” Gebbie said.
“The big question for companies considering deploying this system is: will you recoup the monthly amount of value for your organization?
Managing AI risks
It’s not just the costs of AI that businesses will face next year, but also the growing threat of regulation from governments. “There is certainly an expectation that we will start to see more significant regulatory aspects emerge around generative AI,” Gebbie said.
This combination of rising costs and the threat of heavy regulation means companies may think twice before deploying generative AI. “We definitely see that there will be a slight slowdown in the number of enthusiasts around generative AI next year,” Gebbie added.
Bring back the humans
Along with waning enthusiasm for AI, the analyst firm predicts that by 2026 there will be a revival in demand for human programmers to inject life back into AI-generated code.
“Code augmented or created by AI undoubtedly contributes to the development of innovative solutions,” CCS Insight said in its annual forecast report. “This is also useful for identifying and mitigating software vulnerabilities.”
“However, the human spark of innovation and ingenuity, particularly for non-linear approaches, is becoming a specialty of skilled and experienced programmers, not least because of the lower energy requirements to undertake their work.”