While professionals at every level recognize the benefits of generative AI—citing employee productivity, higher-quality output and cost savings—they are also acknowledging the risks of using the tech in the enterprise, especially related to data privacy and the inability to ensure message consistency. New research from generative AI platform Writer reveals that nearly half of the 450 enterprise executives (directors and above) surveyed believe someone in their company may have inadvertently shared corporate data with ChatGPT.
The adoption of generative AI has been lightning fast: Companies in virtually every industry are investing in (and experimenting with these very new tools—ChatGPT just became public less than six months ago—and rolling them out across every function, including IT, operations, marketing, sales and support, the firm’s inaugural State of Generative AI in the Enterprise report affirms—but could a groundswell of breaches and hacks be waiting around the corner from all of this casual experimental usage?
“Enterprise executives need to take note. There is a real competitive advantage in implementing generative AI across their businesses, but it’s clear there’s a likelihood of security, privacy and brand reputation risks,” said May Habib, Writer CEO and co-founder, in a news release, adding that businesses need to have complete control of their AI— from what data large language models (LLMs) can access to where that data and LLM is hosted. “If you don’t control your generative AI rollout, you certainly can’t control the quality of output or the brand and security risks.”
Other key findings of the research:
Businesses have a love/hate relationship with generative AI
ChatGPT is both the most used and most banned tool: 47 percent of respondents (and 52 percent of respondents in regulated industries) cite using OpenAI’s popular free chatbot at work. CopyAI (35 percent) and Anyword (2 percent) were a distant second and third. But companies banning generative AI in the workplace number almost as high: ChatGPT is the most banned (32 percent), followed by CopyAI (28 percent) and Jasper (23 percent).
All departments are using it
Not only are most organizations using generative AI, but they’re using them across virtually every function. When asked to “select all that apply,” seven functional groups received at least 15 percent of the vote: information technology (30 percent), operations (23 percent), customer success (20 percent), marketing (18 percent), support (16 percent), sales (15 percent) and HR (15 percent).
Generating copy is AI’s top business use
The most often cited use cases of generative AI are creating short copy, such as ads and headlines (31 percent), repurposing existing content for different media/channels (27 percent) and creating long-form content, such as blogs and thought leadership articles (25 percent)—which makes it a very reasonable possibility that a staffer could “over-inform” ChatGPT about something proprietary.
Regardless of risk, the future looks very bright for AI
Generative AI is solidly in the enterprise mainstream, with 59 percent of respondents stating their company has purchased at least one, or plans to purchase one such tool this year. Nearly one-fifth (19 percent) of respondents state their company is currently using five or more generative AI tools. Only 7 percent of respondents say their company isn’t planning to purchase generative AI tools.
Senior leaders see a lot of promise in generative AI. The top benefits they cite are employee productivity (23 percent), higher-quality output (22 percent) and cost savings (20 percent): 56 percent of respondents say generative AI boosts productivity by at least 50 percent, and 26 percent say it’s 75 percent or more.
Download the report here.
Writer conducted a survey of 466 directors and above at organizations spanning industry with over 1,000 employees. The survey was conducted via Pollfish between April 13 and April 15, 2023 on behalf of Writer.