Microsoft is planning to launch a privacy-centric version of ChatGPT in response to concerns that people"s data may be exploited for the training of artificial intelligence models. The move is set to appeal to industries like healthcare, finance, and banking that have refrained from using ChatGPT due to the potential risk of their staff sharing sensitive information with the system.
A separate version of the AI tool will run on dedicated cloud servers where data will be kept separate from that of other customers. The data on these dedicated servers will be isolated from the main ChatGPT system to ensure privacy. The private setup could cost as much as 10 times more than what customers currently pay to use ChatGPT.
OpenAI also plans to launch a new privacy-focused business subscription that won"t feed user data to train models by default. OpenAI has also sold Morgan Stanley a private ChatGPT service using which the bank"s wealth management division is able to ask questions and analyze thousands of the bank"s market research documents. Since Microsoft already has a multi-year, multi-billion dollar investment into OpenAI, it is allowed to resell its products without breaking any terms.
The privacy-centric AI service from Microsoft can be a game changer for companies that tend to deal with important and sensitive data. Samsung banned its employees from using generational AI chatbots at work or on devices they use for work when it found that some of its employees had uploaded company source code. Salespeople from Microsoft are already inquiring organizations about the upcoming product, as many existing customers have contracts with Azure that could prove beneficial in managing the data securely.
Microsoft is expected to launch its private AI service later this quarter.
Source: The Information