ChatGPT
ChatGPT and other Large Lamguage Models (LLMs) are becoming useful to quickly get up to speed on a new scientific topic. Note though, that, unlike a Google Scholar search, which crawls and surfaces information on the web, ChatGPT generates/synthesizes information from the sources on which it has been trained.
Highly specialized and curated scientific information is still largely kept behind publisher paywalls and may not have been included as part of ChatGPT's training. Therefore, the output from ChatGPT may be outdated, incomplete or made up. It is thus up to the user to verify the veracity of the information. If references are provided as part of the output, do check whether the references are real by looking up the source material.
To get the most out of chatting with ChatGPT consider how to best prompt it to perform a specific task. This is called 'prompt engineering'. Google it!
Beware: do not share private or confidential information with ChatGPT as there is no guarantee that the information you volunteered will not be used by the company.
Microsoft Copilot
At present Cornell has access to a protected, walled-off version of Microsoft Copilot. Integrated features include DALL-E.
What is the difference to ChatGPT? This is what Copilot had to say: if you’re looking for a conversational AI chatbot, ChatGPT is a great choice. For programming and code-related tasks, Copilot is more suitable. Combining both tools in your workflow can offer the best of both worlds for personal and professional projects!