As ChatGPT continues to make a significant impact across various industries, it has proven to be an excellent productivity tool for modern sectors such as finance and data. However, concerns have arisen regarding potential data leakage risks, as user-provided content could be utilized by OpenAI for AI model training and later provided to other users.
To address these concerns, Microsoft plans to launch a privacy-centric version of ChatGPT, which will be hosted on Microsoft’s independent GPU servers, isolated from the main ChatGPT system and ensuring no data sharing with other clients. This solution appears to be a cloud-hosted version of ChatGPT, and there are currently no indications that either Microsoft or OpenAI will provide an offline version due to potential core data leakage risks.
Pricing for the cloud-hosted version is rumored to be significantly higher, possibly over ten times that of the standard version. However, as the primary clientele for this service is expected to be in the financial industry, cost should not be a significant barrier.
OpenAI also plans to introduce a privacy-centric business subscription that, by default, will not provide data for model training. A prime example of a typical customer is Morgan Stanley, to whom OpenAI provides a private ChatGPT service capable of analyzing various market research reports and offering insights. It is anticipated that other financial firms will soon invest in this type of business subscription.
Microsoft’s cloud-hosted ChatGPT is expected to arrive later this quarter, while OpenAI’s business subscription should be ready for rollout to large clients soon.