Most pharma and biotech companies ban or restrict the use of ChatGPT by their employees, according to a recent survey. Half of pharma and biotech companies restrict its use, but that jumps to 65% when narrowed down to the top 20 pharma companies, according to the ZoomRx survey.
The chief reason cited by almost all was security and the potential to leak internal data. But the pharma industry isn’t alone in its concerns over the open, generative AI platform. Many leading tech and finance brands, including Apple, Amazon, Citigroup, Goldman Sachs, JPMorgan Chase, Microsoft, Samsung and Verizon, have banned the internal use of ChatGPT.
Andrew Yukawa“What you’re seeing here is a lot of trepidation around putting sensitive company information and sensitive customer information in jeopardy,” said Andrew Yukawa, ZoomRx product manager. “You can imagine a scenario where an employee working with ChatGPT enters some proprietary clinical data or maybe sensitive information collected in a conversation with a colleague. But that action inadvertently adds that information and that data into Open AI’s data training set.”
ZoomRx polled 211 pharma and biotech employees who had from two to 35 years’ experience in the industry and were familiar with at least one generative AI large language model. They interviewed people across a range of jobs, including R&D, sales and marketing, manufacturing, business development, medical affairs and more.
Still, employees may not be too concerned. Some 80% of those surveyed believe AI is currently overrated, and only 25% use ChatGPT more than once a week.
“There’s a large gap in terms of perception versus reality,” Yukawa said. “You’re starting to see this dichotomy between what AI has been presented as in a very theoretical and idealistic sense versus what AI has currently turned out to be in a more practical and tangible sense.”
When asked what was the main reason their company is considering or using any kind of AI, the majority listed reasons like cost reduction, increased efficiency or time savings. Far fewer pointed to business growth or revenue impact use cases, he said.