5 Comments
User's avatar
Elaine Swerdlow's avatar

Thanks, Jason, for this illuminating essay.

Just curious: Will Chat GPT proponents counter by offering to crawl academic sources, rather than just the web, and giving proper attribution?

I still agree with you that so much would still be lost, but just curious.

Elaine

Expand full comment
Jason Steinhauer's avatar

Yes, that's likely to happen at some point. Most academic journal articles are still behind paywalls, controlled by large companies such as Springer-Nature or Elsevier, but at some point there will likely be licensing and cost-sharing agreements that enable that to be crawled. There are also open source databases and library catalogs. Important to remember, though, one of the major points in my book: access to more information does not equate with knowing more about the world. It can, in fact, have the opposite effect.

Expand full comment
Elaine Swerdlow's avatar

Hmm. Thanks, Jason.

Expand full comment
Anil Bahuman's avatar

Are there detection tools that can tell when text has been generated by A.I.?

Thanks for bringing this up and even the failure of current tools. For tools to work 100% of the time, they will have to be incorporated by the AI itself. Chat GPT3 is expected to include this Watermarking technology soon. After that plagiarism tools will work 100%

Expand full comment
Jason Steinhauer's avatar

Thanks, Anil. Appreciate you reading and commenting!

Expand full comment