CFOs should lead adoption of generative AI: Gartner

By Jim Tyson for CFODive.com

CFOs should take a leading part in adoption of generative artificial intelligence rather than solely rely on their IT departments, thereby ensuring the new technology aligns with business strategy and the full range of company operations, according to a Gartner AI analyst.

“CFOs should have a larger role in generative AI than they have in other technologies,” Rajesh Kandaswamy, said Wednesday, emphasizing that recent AI innovations will likely upend the business and operating models at many companies.

A CFO is more familiar than IT staff with a company’s sources of revenue, its biggest costs and its most promising opportunities, Kandaswamy said in an interview at the Gartner CFO and Finance Executive Conference in National Harbor, Maryland.

Assessing generative AI as part of a cross-functional team, “the CFO is probably in the best position to educate the company to start thinking strategically, not the CIO,” he said.

Kandaswamy spoke after leading a discussion by 20 CFOs and finance executives on the opportunities and challenges from generative AI.

Several CFOs hailed the new technology for streamlining tasks such as answering detailed technical questions and writing first drafts of emails or memorandums.

At the same time, the finance executives flagged several risks from the new technology, including the vulnerability of proprietary data, the spread of “deepfake” misinformation, the potential for copyright infringement and the creation of outright fabrications or “hallucinations.”   

When reviewing generative AI content, “always validate it,” Kandaswamy told the CFOs, noting revelations last month that a New York attorney using ChatGPT submitted in court a 10-page brief that cited several bogus court decisions. 

The creators of some generative AI programs have warned about flaws in their algorithms. OpenAI said in March that the accuracy of ChatGPT-4 in business applications is less than 80%. Google says in a disclaimer that “Bard is experimental, and some of its responses may be inaccurate, so double-check information in Bard’s responses.”

Generative AI currently offers a limited, but potentially valuable, set of reliable use cases,  Kandaswamy said.

The technology can provide a starting point for website content, sales memos and marketing material, he said. Companies such as Microsoft and Salesforce have embedded generative AI tools in their software.  

CFOs should consider using generative AI “anywhere they have content generation that is of low risk,” Kandaswamy said.

At the same time, the weaknesses of generative AI — along with its widespread availability — underscore the imperative of constructing safeguards against abuse, he said. “This is a huge moment of democratizing AI.”

CFOs should insist on a policy of “responsible AI,” he said. Such programs ensure that a senior executive is fully accountable for the technology.

Responsible AI guidelines also seek to ensure that the technology serves a broad range of stakeholders; mines high-integrity data; protects against attacks and rogue use; shields user data; avoids harming people, property or the environment; and is explainable, transparent and reliable.

With such safeguards in place, a CFO can confidently encourage experimentation with generative AI, according to Kandaswamy.