Generative AI’s promise of making artificial intelligence more accessible to business users also presents significant challenges for CIOs, who must balance the urgency to capitalize on this transformational technology with the risks of unfettered experimentation across the business.

Research from MIT and others shows early adopters enjoying a variety of benefits from generative AI, including cost and time savings of 70%, effectiveness gains of 30%, and productivity gains up to 50%1. Most of these gains, however, are limited to small proof of concept pilots. The challenge is that because generative AI platforms and tools are easily accessible, these pilots often take place without any governance or IT oversight.

These siloed efforts, which EY Global Consulting AI Leader Dan Diasio calls “shadow AI,” expose enterprises to a variety of risks and challenges, including trust and accuracy, fairness and bias, privacy, compliance, legal, and information security. From a functional standpoint, scalability is also an issue, because while it’s easy to spin up a proof-of-concept, these fragmented efforts lack the infrastructure required to effectively scale and integrate with other initiatives across the enterprise.

Siloed deployment raises another pressing issue: Organizations that become too focused on micro improvements may lack the broader vision of how to leverage generative AI for wholesale transformation.

“People are often thinking about the way a process operates today and how it can be improved, versus thinking about the way they can conduct their work in the future,” Diasio explains. “That means we have organizations moving forward with a large portfolio of really small bets that will not provide the transformative value that generative AI promises to bring.”

Orchestrating a top-down approach to generative AI

Rather than a bottom-up approach to AI initiatives, Diasio recommends that organizations develop generative AI capabilities in an integrated way, from a top-down perspective. This realignment helps drive the requisite culture change, wrests the greatest value from AI investments, and helps to ensure that people are using the ground-breaking technology responsibly.

One of the biggest shifts should be moving beyond a center of excellence (CoE), which organizations often establish to promote experimentation, to a “control tower” paradigm where leadership actively plays a role in determining business cases, orchestrating financial analysis, capturing institutional knowledge, and retraining the workforce for new ways of working. “A control tower is more top down and works on three or four initiatives, versus a CoE, which often prioritizes a list of 100 or 200 use cases,” Diasio explains.

CIOs can play a critical role in this top-down reorientation by taking the following steps:

Foster alignment among senior leadership on value. CIOs should lead the conversation, educating peers on the technology and building consensus on strategy at the executive and board level. CIOs should strive to embed technology into the business strategy, creating a “transformation portfolio” rather than an “AI portfolio,” and steer the organization away from “shiny object syndrome,” featuring an overwhelming number of use cases but little to no value.

Lean into governance. Create an AI council that can lead the responsible AI agenda, continuously evaluating ethics controls, privacy impact assessments, and business impact assessments, and use those findings to recalibrate the AI systems. CIOs should also lead efforts to document AI and capture the inventory of AI systems in use across the company or by third parties2.

Define the technology infrastructure required to activate AI adoption. Most organizations do not have a technology or data platform that is ready to integrate their data with generative AI – the data lake or warehouse is missing a critical layer. Organizations will need to invest to capture knowledge from across the business and serve up their data to be searchable to customize large language models (LLMs).

Build out training and awareness. CIOs play an important role in creating the infrastructure for knowledge to be collected across the organization and to build programs that drive awareness of what generative AI is capable of. CIOs should work with HR and other business peers to build a professional development plan that includes AI-related training programs, career paths, and a way to reward new AI skillsets.

Whether leading the charge or facilitating critical building blocks, CIOs play a key role in orchestrating a top-down approach to generative AI. This shift in both mindset and structure will help to ensure the technology delivers truly transformative benefits across the organization.

Learn more about how to build confidence in AI to create exponential value for positive human impact.

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.

1“GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models,” March 2023, OpenAI, OpenResearch, University of Pennsylvania

2 EU AI Act readiness, EY