This is great. I'm still relatively new to librarianship and understand the responsibility to be critically skeptical and to stress test "new" tech, but I don't understand this pattern of discarding new tools before fully understanding their potential utility. Baby with the bathwater and all that.
I don't think there should be much pushback. The antagonism to Google Scholar never made sense to me, since once students complete their degrees and go out to the "real world" they will NOT have any access to ProQuest, WoS, etc., so it always made sense to me to teach them how to use Google Scholar and now using AI for search. In terms of finding the open access stuff "in the real world" that is just what people are gonna do, whether we like it or not. At the end of the day, for librarians, this is an opportunity for information literacy in terms of teaching how to identify sources, understand what sources are, identify debates, rhetorical citations................there is a lot of good stuff for us to do in terms of information literacy thanks to gen AI.
They are several reasons why librarians might be biased against Google Scholar and favour tools they subscribe to
1. For many librarians they need to show the tools they subscribe to has high usage. If users use Google Scholar to search over say Scopus or some Ebsco databases that means potential usage lost and/or a sign their judgment is wrong
2. Also never underestimate the ease of having vendor support at your back at all. In the mid 2010s, researchers kept asking us to teach the bibliometrics team to teach Google scholar. But the leader of the team who was supposed to be a "expert in bibliometrics" refused repeatedly and later I realised it was because unlike Scopus/Web of Science where you always had a safety harness of asking support, Google Scholar famously does not answer individual user queries (though they used to have a team that engaged with repositories), so she lacked confidence to even talk about it.
nice. In addition, I'd add how AI tool producers have a choice:
1)Do they cover up their flaws, make certain aspects ambiguous in their documentation, market their product as AI search, and avoid admitting it's not the ideal AI search and really just a boolean constructer, for example. And you only discover this from testing, asking them direct questions to draw it out, or reading Aaron Tay.
2)Or do they admit up front the status of their AI offering, yes, these are our products currently, we don't offer the ideal AI search as affecting the environment is unavoidable if you do, we are exploring possibilities for the future like "chunking". It is a dynamic space where things are constantly changing.
As a potential customer, the second option earns much more trust than the first.
1. I know your "boolean constructor" is just an example but in this particular case, I think vendors were extremely upfront they were doing that. Especially since they knew many librarians loved that paradigm because it's easy to understand.
This is great. I'm still relatively new to librarianship and understand the responsibility to be critically skeptical and to stress test "new" tech, but I don't understand this pattern of discarding new tools before fully understanding their potential utility. Baby with the bathwater and all that.
So many good points. Thanks for writing this.
Thanks. I am expecting a lot of pushback….
I don't think there should be much pushback. The antagonism to Google Scholar never made sense to me, since once students complete their degrees and go out to the "real world" they will NOT have any access to ProQuest, WoS, etc., so it always made sense to me to teach them how to use Google Scholar and now using AI for search. In terms of finding the open access stuff "in the real world" that is just what people are gonna do, whether we like it or not. At the end of the day, for librarians, this is an opportunity for information literacy in terms of teaching how to identify sources, understand what sources are, identify debates, rhetorical citations................there is a lot of good stuff for us to do in terms of information literacy thanks to gen AI.
They are several reasons why librarians might be biased against Google Scholar and favour tools they subscribe to
1. For many librarians they need to show the tools they subscribe to has high usage. If users use Google Scholar to search over say Scopus or some Ebsco databases that means potential usage lost and/or a sign their judgment is wrong
2. Also never underestimate the ease of having vendor support at your back at all. In the mid 2010s, researchers kept asking us to teach the bibliometrics team to teach Google scholar. But the leader of the team who was supposed to be a "expert in bibliometrics" refused repeatedly and later I realised it was because unlike Scopus/Web of Science where you always had a safety harness of asking support, Google Scholar famously does not answer individual user queries (though they used to have a team that engaged with repositories), so she lacked confidence to even talk about it.
nice. In addition, I'd add how AI tool producers have a choice:
1)Do they cover up their flaws, make certain aspects ambiguous in their documentation, market their product as AI search, and avoid admitting it's not the ideal AI search and really just a boolean constructer, for example. And you only discover this from testing, asking them direct questions to draw it out, or reading Aaron Tay.
2)Or do they admit up front the status of their AI offering, yes, these are our products currently, we don't offer the ideal AI search as affecting the environment is unavoidable if you do, we are exploring possibilities for the future like "chunking". It is a dynamic space where things are constantly changing.
As a potential customer, the second option earns much more trust than the first.
1. I know your "boolean constructor" is just an example but in this particular case, I think vendors were extremely upfront they were doing that. Especially since they knew many librarians loved that paradigm because it's easy to understand.