A survey of 600 chief data officers published today finds nearly half (45%) half have implemented generative artificial intelligence (AI) to some degree, but nearly all (99%) have encountered some type of roadblocks.
Conducted by Informatica, the survey finds the top issues cited by adopters are data quality (42%), data privacy and governance (40%) and AI ethics (38%).
The survey makes it clear that data management issues are at the heart of these challenges, says Jitesh Ghai, chief product officer at Informatica. “Garbage in, garbage out is still perennially the challenge,” he adds.
Overall, the survey finds 53% of the respondents that work for organizations that have yet to embrace generative AI expect they will, with 36% planning to do so within the next two years.
However, a full 98% also acknowledged organizational obstacles are holding back efforts, including a lack of leadership support (45%), inability to justify return on investment to gain budget approval (45%), and lack of cooperation/alignment across business units (44%).
Ultimately, nearly three quarters of respondents (73%) are or plan to employ generative AI to improve time to value with faster data insights, while two-thirds (66%) also want to drive more productivity through automation and augmentation.
The challenge is 58% of respondents are already using five or more data management tools. A total of 41% are struggling to balance more than 1,000 data sources.
Not surprisingly, the ability to deliver reliable and consistent data fit for generative AI (39%) and improving data-driven culture and data literacy (39%) are the top data strategy priorities in 2024. As a result, data privacy and protection (45%), data quality and observability (41%), and data integration and engineering (37%) are all top areas of investment in the coming year.
It’s not clear to what degree organizations will benefit from generative AI investments in 2024, but nearly all to varying degrees will be trying to find a way to operationalize it across multiple processes. There are plenty of low-hanging fruit opportunities to, for example, improve customer service. The challenge will be identifying use cases that provide organizations with a sustainable competitive advantage. Generative AI is rapidly becoming the new table stakes required to simply remain competitive.
Regardless of the use case, however, the one thing that is certain is generative AI initiatives will expose lax data management practices that are all too common with many enterprise IT organizations. Much of that data is either conflicting or flat out wrong so while much of the funding for generative AI initiatives will need to be applied to improving data quality.
In the meantime, business and IT leaders should assume that, in addition to the volume of data that needs to be managed and continuing to exponentially grow, there will soon be a proliferation of the large language models (LLMs) being used to enable generative AI. Those LLMs will all need to be trained using data that has been well vetted because, otherwise, the outcome generated by those models is all too often going to be anything but reliable.