OpenAI addresses corporate data leakage jitters with ChatGPT Enterprise

OpenAI launched a special version of its ChatGPT AI chatbot for use within businesses. The new bot, called ChatGPT Enterprise, can help workers craft clearer communications, speed up coding tasks, explore answers to complex business questions, and assist with creative work, OpenAI said in a Monday blog post.

OpenAI believes that over 80% of Fortune 500 companies’ teams already use the consumer version of ChatGPT, and it has heard from at-work users that they’d like a version of the chatbot with commercial-grade security features.

“We, of course, get a lot of pull from companies that are either actively building their own version of ChatGPT, because they want a version of ChatGPT for their company, or are basically asking us day in and day out ‘when are you building it?’” OpenAI chief operating officer Brad Lightcap tells Fast Company.

Some corporate leaders have shied away from API-based services that run on OpenAI’s servers because they fear their proprietary data might leak out into the world. Lightcap says addressing those concerns is core to the enterprise product’s development.

“It’s just been a [process of] making sure that it’s secure and private from a data perspective,” he says. “We, of course, don’t train [our models] on any of this data; we retain it because it’s required to run the product, but we don’t look at any of that data. Enterprises are entirely in control of it.” 

OpenAI says companies like Block, Canva, Carlyle, The Estée Lauder Companies, PwC, and Zapier have been trying out ChatGPT Enterprise for a few months.

Lightcap says ChatGPT Enterprise lets large companies use the GPT-4 model “with low latency, message caps, and ultimately with customizability so that you can make ChatGPT really personal to your enterprise.” 

He adds that OpenAI will probably release a version of the chatbot for smaller businesses called ChatGPT Business, but didn’t provide further details.