We don’t have a uniform position on this; individual team members across GiveWell have found where AI is and is not useful to them in their work. Where there are legal constraints regarding AI use, we do comply with them, but that’s really the minority of use cases. Overall I’d say our approach has been to try to use AI tools where they have the potential to make human work better; we have not generally found them as useful as a replacement for human thought and communication. For example, on our Grants Admin team we have been experimenting with using AI tools to transfer information between systems that (for whatever reason) can’t talk directly to each other. But we don’t (and likely wouldn’t) use AI tools to read & respond to correspondence from our grantee partners.
We don’t have a uniform position on this; individual team members across GiveWell have found where AI is and is not useful to them in their work. Where there are legal constraints regarding AI use, we do comply with them, but that’s really the minority of use cases. Overall I’d say our approach has been to try to use AI tools where they have the potential to make human work better; we have not generally found them as useful as a replacement for human thought and communication. For example, on our Grants Admin team we have been experimenting with using AI tools to transfer information between systems that (for whatever reason) can’t talk directly to each other. But we don’t (and likely wouldn’t) use AI tools to read & respond to correspondence from our grantee partners.