It has been reported that Microsoft is developing a fresh iteration of ChatGPT, specifically designed for business customers with a focus on privacy. Microsoft has already integrated the GPT-4 AI technology, which powers the large language model of ChatGPT, into Bing Chat to provide users with a similar chatbot experience.
According to The Information, Microsoft’s agreement worth $10 billion with OpenAI, the creators of ChatGPT, grants Microsoft the authority to provide modified versions of ChatGPT to specific users and organizations. The tech giant is presently developing a privacy-centric version of ChatGPT that would function via dedicated servers.
The privacy-centric version of ChatGPT developed by Microsoft will utilize dedicated Azure cloud servers exclusively for business purposes. The data stored on the isolated servers will remain completely separate from the main ChatGPT system, ensuring complete privacy. However, this bespoke service comes at a high cost, and enterprise customers will need to pay a substantial sum to avail themselves of its benefits. According to the report, the price of this service could be as much as ten times higher than the current ChatGPT Plus subscription fee of $20.
Microsoft’s decision to create a ChatGPT exclusively for internal use could appeal to numerous major technology and banking companies that prohibit their workers from utilizing public or free chatbots. Samsung, for example, supposedly prohibited ChatGPT on its corporate servers because certain employees were discovered to have shared confidential information with the chatbot, which could pose a significant privacy risk in the future.
ChatGPT, which places a strong emphasis on privacy, is expected to launch at some point during the current quarter. Additionally, OpenAI, the organization behind ChatGPT, has indicated that it is considering providing a comparable service to enterprise clients. In an interview with CNBC last week, the CEO of OpenAI revealed that the company had temporarily halted the use of customer data from paying clients to train its large-language models, including GPT.
According to the report which cites Altman, “Our customers have made it clear that they do not want us to use their data for training purposes. As a result, we have revised our plans and will refrain from doing so.” Nonetheless, OpenAI continues to enhance its chatbot using sources other than APIs. This implies that the language model is being trained and refined based on inquiries posed by individuals on the publicly accessible version of ChatGPT.
The US government has expressed concerns about privacy, prompting a meeting between US Vice President Kamala Harris and high-level executives from Microsoft, Google, OpenAI, and Anthropic to discuss the risks associated with AI. During the meeting, Harris emphasized the “ethical, moral, and legal responsibility” of private companies to ensure the safety and security of their AI products, according to a statement obtained by Reuters.
According to the US Vice President, chatbots powered by AI, such as Google’s Bard, Microsoft’s Bing Chat, and OpenAI’s ChatGPT, could be both beneficial and detrimental to their users.