Shieldbase
Aug 8, 2024
IT leaders are navigating the complexities of generative AI in real-time, facing challenges ranging from talent shortages to the intricacies of data management.
The rapid adoption of generative AI has sparked a wide range of emotions within organizations, from excitement and anticipation to anxiety and concern.
A global McKinsey survey conducted in May revealed that 65% of organizations are now using generative AI—almost double the percentage from 10 months earlier. As adoption grows, so do the use cases.
As seen with previous technological advancements, many organizations are initially applying generative AI to areas where it can deliver tactical benefits, such as streamlining existing processes and reducing costs, notes Jim Rowan, a principal at Deloitte Consulting. This strategy allows organizations to capture quick wins while building familiarity and confidence with the new technology.
However, companies are at different stages in their generative AI journey. Some early adopters are scaling pilot projects to production by integrating multiple use cases, while others are focusing on proofs of concept or adopting AI technology embedded in third-party software. Still, others are taking a wait-and-see approach.
As organizations transition from learning about generative AI to piloting and deploying full-scale implementations, they are encountering six key challenges.
Tech Talent Shortage as a Major Hurdle
Tech talent remains the top barrier to AI adoption. Organizations with high levels of AI expertise (33%) tend to view generative AI more positively, but they also feel increased pressure to adopt the technology, perceiving it as a potential threat to their business models, according to a Deloitte report on generative AI. This has prompted IT leaders, even those with experienced AI teams, to reevaluate their talent strategies, with upskilling emerging as a critical component in addressing AI skills gaps.
The Elusive Bottom-Line Impact of Generative AI
While the initial excitement around generative AI has waned, few projects are currently delivering measurable bottom-line impact, according to Aamer Baig, a senior partner at McKinsey. Only 15% of surveyed companies have a clear path to earnings improvements from their generative AI initiatives. Baig advises organizations to concentrate on initiatives that solve real business problems, are technologically feasible, and carry minimal risk. A Deloitte report found that 48% of organizations do not expect to see transformational benefits from generative AI for another one to three years.
Legal and Regulatory Uncertainty
Legal and regulatory uncertainties are slowing down the deployment of generative AI platforms at scale. The high cost of being an early adopter, coupled with concerns about legal liability, has made many organizations hesitant to be first movers in this space. A Deloitte study found that compliance (28%) and governance issues (27%) are significant barriers to AI adoption, with less than half of respondents (42%) feeling they have adequately governed generative AI adoption and mitigated its risks.
Managing Costs Amid High Expectations
The cost of implementing generative AI is a major concern for organizations. The technology's high compute intensity, along with the changes it necessitates in workflows, business processes, and KPIs, requires substantial investment. Organizations must also budget for risk management, hallucination training, and ongoing maintenance. The scarcity and cost of hardware, power, and data for training models further complicate cost management.
Data Challenges in Generative AI
Obtaining high-quality data remains a significant challenge for organizations using generative AI. The models require vast amounts of accurate and relevant data to function effectively. McKinsey's Baig advises organizations to focus on data that supports multiple use cases rather than striving for perfection, which can be overwhelming.
Generative AI is Here to Stay
Despite the challenges, IT leaders recognize that generative AI is not a passing trend. Michael Corrigan, CIO of World Insurance, notes that while generative AI is powerful and evolving rapidly, it is also maturing slowly. A strategic approach and careful implementation are required to harness its full potential.
As organizations grapple with these challenges, the consensus among IT leaders is clear: generative AI holds significant potential, but it must be implemented thoughtfully, securely, and at scale. This will require ongoing investment, a deep understanding of the technology, and strong leadership from CIOs to guide their organizations through this transformative period.