Regardless of the numerous issues round generative AI, companies are persevering with to discover the expertise and put it into manufacturing, the 2025 AI and Information Management Government Benchmark Survey revealed.

Generative AI is more and more being put into manufacturing, survey creator Randy Bean and Tom Davenport, professor of IT and administration at Babson School, noticed in this year‘s study. Final 12 months, solely 5% of respondents stated they’d put the expertise into manufacturing at scale; this 12 months 24% have achieved so. Early-stage manufacturing has additionally elevated, from 25% to 47%. Solely 29% are nonetheless simply experimenting with generative AI, versus 70% within the 2024 examine.

The examine examined AI and generative AI utilization in 125 Fortune 1000 organizations. Chief knowledge officers (CDOs), chief knowledge and analytics officers (CDAOs), and chief synthetic intelligence officers (CAIO) collectively made up 91% of survey respondents; 4% held the title of CIO or CTO, and three% have been C-suite execuitves. Most (85%) have been in North America.

These knowledge and AI leaders, Bean and Davenport concluded, are constructive about generative AI’s enterprise worth, with most — 58% — feeling that the first worth is coming from productiveness positive aspects or efficiencies.

And regardless of acknowledging enterprise threats posed by AI, together with the unfold of disinformation or misinformation (53.2%), moral bias (19.8%), and job loss or displacement (4.8%), 96.6% of these surveyed view the general influence of AI as useful.

Who runs AI?

As for the position a CIO ought to play in working AI, Bean stated in interview with CIO that “that is the place Tom and I’ll have slightly little bit of variance in perspective.”

There may be, he stated, an actual divergence in organizations as as to if an AI initiative “sits on the enterprise facet or the expertise facet. I imply, traditionally, it sat on the expertise facet, however there’s been a development, a migration to an increasing number of knowledge and AI leaders sitting on the enterprise facet of it. Personally, I consider it ought to sit on the enterprise facet.”

“[We have] a distinction of opinion as a result of he thinks ‘oh, the info individual ought to be a enterprise individual, and never report back to the CIO,’” Davenport stated. “If a CIO is kind of centered on enterprise transformation, I feel that it’s much better to have all these subsidiary features reporting to her or him.”

However, he added, “in the event you’re the pinnacle of selling or one thing, you may have a expertise initiative you need, and there are 4 or 5 totally different tech leaders that it’s a must to go to determine tips on how to get achieved, what that you must get achieved, that’s very irritating, I feel, and results in an entire number of issues.”

Survey respondents have been equally divided, with 36.3% reporting that knowledge and AI operations report back to the enterprise, 47.2% saying expertise management holds the reins, and 16.5% saying they produce other reporting relationships, akin to to transformation management.

What’s the enterprise worth?

The survey additionally discovered that few corporations are measuring productiveness positive aspects from AI fastidiously. Bean stated he was not stunned, as a result of reality “that 90% of Fortune 1000 corporations are legacy corporations, that means they’re principally a technology or older. These corporations are typically extra cautious and danger averse on account of giant buyer franchises constructed over generations and even centuries or extra.”

He added, “in spite the entire enthusiasm I used to be listening to within the trade, every time I met with the chief knowledge officers, the chief AI officers, or any of those corporations, to an individual, they stunning a lot stated, ‘you understand, we’re at an early stage.’”

Requested if something from this 12 months’s survey stunned him, Davenport stated that final 12 months, generative AI “actually gave the impression to be liable for enormous adjustments within the knowledge cultures of group, nearly a doubling, or greater than a doubling, really, within the proportion of corporations that stated that ‘we’re knowledge pushed, we’ve an information pushed tradition.’ That fell again to about half of the rise this 12 months. I feel what it begins to recommend is generative AI, by itself, shouldn’t be going to single handedly rework the info tradition of organizations.”

The report famous that knowledge and AI leaders surveyed are “additionally constructive about genAI’s enterprise worth. Virtually half, 46%, say that the enterprise worth is both already excessive or vital and rising; one other 32% say it’s ‘modest however growing. Most — 58% — really feel that the first worth is coming from “exponential productiveness positive aspects or efficiencies.”

Nonetheless, the authors famous in an article in regards to the outcomes revealed within the MIT Sloan Administration Evaluation, “corporations shouldn’t take such confidence on religion. Only a few corporations are literally measuring productiveness positive aspects fastidiously or determining what the liberated data employees are doing with their freed-up time. Just a few educational research have measured genAI productiveness positive aspects, and after they have, they’ve usually discovered some enhancements, however not exponential ones.”

There are nonetheless knowledge and AI management challenges to take care of too. The report famous that the CDO position continues “to be very a lot a piece in progress as organizational knowledge and AI wants quickly evolve. The CDO position is characterised by excessive turnover, brief tenures and never being nicely understood. Whereas CDOs face headwinds as organizational change brokers, most consider that the position is evolving in the fitting path.”

However regardless of these challenges, the report stated, “most organizations [89%] consider that AI would be the transformational expertise in a technology, akin to the web.”