Enterprises enthusiastic about generative AI, Foundry survey shows

There are tensions between IT and line-of-business leaders about the risks of moving too fast or losing control of data, however.

Generative AI is already making deep inroads into the enterprise, but not always under IT department control, according to a recent survey of business and IT leaders by Foundry, publisher of CIO.com. The survey found tension between business leaders seeking competitive advantage, and IT leaders wanting to limit risks.

Some 62% of respondents said their organizations were actively using generative AI, with 23% saying they were in the early stages of exploring its use, and 14% saying they were considering implementing it. That leaves just 1% that has either checked out generative AI and dismissed it, or have no plans to use it at all.

Interestingly, non-IT leaders were more likely to report actively using generative AI (73%) than IT leaders (59%), suggesting there’s plenty of experimentation going on beyond the purview of the IT department. Enterprises with 5,000 or more employees were more likely (69%) to be trying the technology than smaller ones (57%).

Enthusiasm for generative AI technology, a branch of AI that can be used to autonomously create new content such as images, videos, or text, varied from one industry to another. Those in retail or financial services were most likely (62%) to report active use of it, closely followed by manufacturing, production, and distribution (59%), and technology (56%).

Tension at the top

There were some interesting splits when Foundry asked respondents whether the greater risk to their organization was moving too fast (adopting generative AI despite security risks or ethical issues) or too slow (being seen as a laggard, or not creating competitive advantage).

IT leaders were more likely (56%) to see moving too fast as the greater threat, while moving too slow was seen as more dangerous by non-IT leaders (52%).

Financial services industry respondents were evenly split on the question. Retailers are definitely risk-takers, with 60% seeing moving too slow as the bigger threat, while those in technology prefer to take precautions, with 58% saying moving too fast is more dangerous.

Furthermore, companies with 5,000 employees or more were most likely to be cautious (75% saying moving too fast is the bigger threat) and smaller companies see slowness as the greatest risk (62.8%).

Brakes on generative AI use

The thing most likely to curb enthusiasm for generative AI (41% overall, and 42% among IT leaders) was the law—or at least legal issues related to generative AI output. These issues could include copyright concerns, privacy violations, or liability for providing or acting on hallucinations of a poorly trained or implemented AI system.

Non-IT leaders, however, were most concerned (40%) about loss of control over corporate data. This only preoccupied 23% of IT leaders, perhaps because they better understand the technological options available for data loss prevention, or for training and running generative AI models on-premises or in private clouds to keep corporate data safe.

Concern about employees going rogue and using generative AI without permission was higher among IT leaders (22%) than among non-IT leaders (14%). That could be because non-IT leaders get to see the benefits of early and unsanctioned experimentation, while IT leaders then have to clean up.

One of generative AI’s superpowers is making mistakes faster, but only 13% of survey respondents (14% in IT, 9% non-IT) cited a lack of confidence in generative AI results as their top concern.

How enterprises use generative AI

Almost 90% of survey respondents said they have generative AI projects of one kind or another underway or have just started them.

Training and upskilling employees on generative AI was the area in which they were most advanced (51% underway, 38% just started), with financial services companies the most likely (59%) to have training projects already underway. Getting generative AI tools in users’ hands came a close second (50% underway, 38% just started). Both those areas could be considered to relate to pure technology, and thus why IT leaders were more likely than non-IT leaders to have projects already underway.

However, non-IT leaders were far more likely to have started work on establishing generative AI policies and guidelines than IT leaders (65% vs 42%), or on identifying use cases (59% vs 38%), the survey found. The industry most likely to have already identified use cases was retail (49%), ahead of tech and manufacturing (42%) with financial services (32%) in last place.

Generative AI generates spending

Asked in which areas their organization would be making AI investments this year, only one respondent (an IT leader in a manufacturing, production, or distribution industry) said “nowhere.”

Spending on AI-enabled applications is on the cards for 76% of respondents, while 68% plan to add headcount for AI-related roles such as data scientists or prompt engineers. Also, 68% will increase security spending, 55% will boost AI-related cloud spending, and 51% will upgrade infrastructure to run AI workloads.

Smaller companies see the greatest need to spend on AI-related security (72% vs 63%).

There were some discrepancies in the spending plans of IT and non-IT leaders, too, with IT seeing the greatest need for spending on AI-enabled applications (79% vs 65%), AI-related headcount (69% vs 63%), and security (70% vs 57%). Non-IT leaders were more likely to want to spend on cloud services (69% vs 51%) and infrastructure upgrades (59% vs 49%).

Foundry conducted its survey in early July, asking senior executives about their preparations for and use of generative AI. Of the 447 respondents, 90% of respondents held C-level roles (CEO, CIO, CTO, CSO, CISO). Other respondents were managers, directors, or VPs. The companies operated primarily in manufacturing, production, distribution, retail, or financial services and the median number of employees in their organizations was 3,750.