Nearly Half of Firms Are Drafting Policies on ChatGPT Use

Nearly half of human resource leaders polled by consulting firm Gartner said they’re in the process of formulating guidance on employees’ use of OpenAI’s artificial intelligence chatbot ChatGPT.

What those policies will look like may end up varying widely. Some Wall Street firms, like Bank of America Corp. and Goldman Sachs Group Inc., have banned the chatbot, while hedge fund giant Citadel has embraced it.

At the same time, one-third of HR leaders surveyed by Gartner said they aren’t planning to issue any policies on employees’ use of ChatGPT, even as experts raise concerns about copyright infringement and data privacy, and caution users against the chatbot’s tendency to, at times, simply make stuff up.

Already, over 40% of professionals polled by Fishbowl, a social platform owned by employer review site Glassdoor, have used ChatGPT at work. Software developers, consultants and bankers are among the early adopters who have used the tool to write emails, reports and bits of code. Most went rogue, according to the Fishbowl survey, experimenting with the tool without telling their bosses.

Wall Street firms have started to crack down. Alongside Bank of America and Goldman Sachs, Citigroup Inc., Deutsche Bank AG and Wells Fargo & Co. have banned the use of ChatGPT. But they’re in the minority: So far, only 3% of the HR leaders surveyed by Gartner said they’ve banned ChatGPT for any business purpose.

Others, like Citadel, are taking the opposite tack, negotiating an enterprise-wide license for the tool.

“This branch of technology has real impact on our business,” Ken Griffin, its billionaire founder told Bloomberg, “everything from helping our developers write better code to translating software between languages to analyze various types of information that we analyze in the ordinary course of our business.”

Meanwhile, Microsoft Corp. debuted its revamped suite of Office applications on Thursday, integrating OpenAI’s new GPT-4 AI model into Excel, PowerPoint, Outlook and Word. The software is currently being tested with 20 companies, including eight in the Fortune 500 that Microsoft declined to name.

The firms working on rules around the use of generative AI are likely still in an exploratory phase, according Eser Rizaoglu, senior director analyst in the Gartner HR practice.

“They're probably questioning how much guidance, which roles will potentially use it or will not be able to use it, and if they should completely ban it or not,” Rizaoglu said. “A lot of leaders are working with IT, legal, compliance and auditing to understand: What are the risks, what are the potential impacts? And then how do we take an approach accordingly?”

At the same time, one-third of HR leaders polled by Gartner said they aren’t planning to issue any policies on employees’ use of ChatGPT. Rizaoglu said that might be because the new technology may be irrelevant to their organization and the industry they’re in, or they believe it’s just a passing trend. Another group may think the responsibility for providing guidance instead lies with the legal or IT departments.

Others polled by Gartner have taken a middle-of-the-road approach, neither banning nor ignoring the chatbot, but warning staff that the chatbot’s answers aren’t necessarily reliable or confidential, and can be analyzed to see whether they’re AI-generated.

Still, given the risks to accuracy, data security and privacy, “the diligent thing would be to assess what the potential risks are for the organization and put in some guidance accordingly to ensure that the organization is mitigating any risks that could occur later on,” Rizaoglu said.