News · November 30, 2022
OpenAI opens ChatGPT as a free research preview
OpenAI released a conversational model that answers questions, writes text and helps with code in a chat interface. It is free to try, and its limits are as important as its abilities.
What happened
- OpenAI introduced ChatGPT on November 30, 2022 as a conversational model and a sibling of InstructGPT, free to use during a research preview.
- According to OpenAI, ChatGPT was fine-tuned from a GPT-3.5 series model that finished training in early 2022, using reinforcement learning from human feedback, and was trained on Azure AI supercomputing infrastructure.
- OpenAI listed limitations: it can give plausible-sounding but wrong answers, is sensitive to how a question is phrased, is often verbose, usually guesses intent instead of asking, and may respond to harmful instructions; the company said it uses its Moderation API to warn about or block some unsafe content.
- SiliconANGLE reported that in an OpenAI demo the system asked for more detail about a faulty code snippet and then explained the bug and suggested a fix.
Why it matters
The chat format lowers the barrier: anyone on a team can now try a large language model without an API key. Expect engineers, marketers and support staff to start using it this week, with or without a policy.
So write the policy now: no confidential code, customer data or credentials in prompts, every output reviewed by someone who can judge it, and no published claims that have not been checked. Within those rules, it is worth testing for first drafts of documentation, ad copy variants and explanations of unfamiliar code.
Sources
- OpenAI: Introducing ChatGPT
- SiliconANGLE: OpenAI debuts new conversational AI system
This post is CMM’s own summary and commentary on publicly reported news. Facts are attributed to the sources linked above; please read them for full details. CMM is not affiliated with these publishers or with the companies mentioned, and trademarks belong to their owners. Prepared with AI assistance. Corrections or requests: [email protected]. See our editorial policy.
Want help with this?
Tell us what you need, and we will suggest the right team model on a short call.
Talk to us