Hacker News Reader

Comment by ChrisMarshallNY | Hacker News Reader

This is one of those "it depends" areas.

For example, if there is a specific advantage to using fluent English text, and the submitters are not fluent in English, then using ChatGPT seems like a fairly natural way to handle that. I'll bet that Google Translate is already being used for it.

Not sure that AI is quite up to actually falsifying the research (although I'll bet it will allow fakers to do much more realistic fake data).

View on HN

Yup. I'm not a native English speaker, thus I use chatgpt (and before this, quillbot or grammarly) to rephrase my sentences to become clearer, concise and have that 'professional' tone, apart from make them grammatically correct.

Even native English speakers could use it.

For example, I am constantly reading prose by extremely intelligent and well-educated people, that is almost painful to read.

One of the things that ChatGPT is great for, is generating eminently readable prose. In fact, I'm recommending to my team, that we consider using it to generate some of our help dialogs.

In many cases, I wish more folks would use it.

Whatever definition of “falsifying research” you have, AI is up to it. The best illustration is the lawyer earlier this year that assumed ChatGPt was truthful and used nonexistent references with complete text.

Fair point, but we'll see. A lot of folks will get away with it, but some folks will hose it up (like the lawyer, who may end up getting disbarred, I think).

Reply on news.ycombinator.com