OpenAI allows companies to add more of their own data to the GPT-3.5 Turbo chatbot and thereby fine-tune the tool based on their needs. OpenAI claims that the resulting custom model can match or exceed the capabilities of GPT-4 for certain tasks. This fall, it is planned to open GPT-4 for the same purpose.
The fine-tuning essentially allows companies to make ChatGPT into a more targeted model. Guided training creates a unique bot for the company that provides, for example, reliable answers in a specific language or more concise wording. Until now, business customers were limited to GPT-3 variants.
The model is pre-trained until September 2021 before inputting company data. According to OpenAI, none of this data is used to train models outside of the customer’s company.
As part of its AI Builder and Power Virtual Agents services, Microsoft also offers enhanceable GPT-based models designed to build responses based on internal enterprise data. Microsoft provides them to summarize information or create content for email campaigns. Like OpenAI’s fine-tuning bots, Microsoft’s customizable AI bots can connect to company data to create responses based on individual business insights.
Source: The Verge
