I don’t know if they have accepted a protocol for summering with AI on this site but on OpenAI Discourse froum we are
- adding a standard tag to the summary for easier finding
- noting the LLM make, model and version
- noting the date the summary was created.
The standard tag is Summary created by AI.
Example search.
The details with each summary at the bottom
Summarized with AI on Aug 01 2023
AI used: gpt-4-32k
Yes this is all done by hand now but summaries are not actually created that often except if someone gets an inkling to add them when needed when the option is first available.
The reason for the standard tag is that I suspect sometime in the future, an HTML tag or such might be created to identify content that was created with AI so that it can be excluded for training sets or such. When that happens it will be easier to identify the content that needs the tag retrofitted.
I know wishful thinking but something similar is really needed; I don’t see the AI as a fad that will just pass.
A specific case of why AI generated information needs to be identified.
When demonstrating some of the uses of ChatGPT with regards to the programming language Prolog a completion hallucinated the name Rodney Byrd when the person’s real name is Lawrence Byrd The hallucination needs to be kept out of search engines, training sets, etc.