“We Couldn’t Generate an Answer for your Question”

Editor’s note: We welcome a guest blog post from Jay Singley, Document Delivery and Circulation Desk Manager at North Carolina School of Science and Mathematics.

In March 2025, Ex Libris unveiled their AI-powered Research Assistant tool for institutions using Summon. Within a week, Summon users reported error messages with specific search terms and topics. The first error reports shared in a listserv for Summon users contained “Tulsa race riot” and “Tulsa race massacre.” Test searches by reference librarians and systems librarians containing these search terms generated no results.

I work in an academic library in a public state university system serving high school students and have since turned on Research Assistant in our Summon Preview Environment (a beta testing environment hidden from users). I have begun testing terms and topics systematically using a methodology akin to Matthew Reidsma’s auditing of algorithms, which seeks to expose the implicit biases of opaque technological systems.

In addition to “Tulsa race riot” and “Tulsa race massacre,” my audit of “controversial” search terms conducted in March-April 2025 turned up error results for the following:

Genocide in Palestine
Gaza war
Rwandan genocide
Armenian genocide
Genocides across the world
History of genocides
lynching
lynching in the united states
lynchings in the united states
january 6
covid
covid data
COVID-19

Ex Libris responded to initial reports about “Tulsa race riot” and “Tulsa race massacre” error results with the following:

“This is due to safeguard policies enforced by our AI service provider to support ethical and responsible AI use. It’s not something the Primo or Summon applications control.

If a query does not return a result, please try different phrasing. For example, ‘tulsa black wall street’ will return results that are directly related to the race riots.”

Despite requests for more information in the same listserv, Ex Libris has not provided a comprehensive response to librarians’ concerns about who the third-party provider is, what the ethical and responsible AI use policies are, whether local control can be made available, what known search terms and topics are obstructed, and more. As one librarian aptly noted, academic users want and need access to information that may be blocked by AI use policies meant for the general public.

An essential note is that Summon’s Research Assistant is not limited to an institution’s catalog. Instead, the AI tool searches the entire Central Discovery Index of Ex Libris’ available records regardless of whether the material is actually available to the user or not. (As of June 2025, Ex Libris has provided an option to limit search results to a user’s institutional catalog.) When the AI generates an error message (or perhaps more aptly refuses to run a search or share results), it is not necessarily because the material is unavailable. Rather, the third party “safeguard” policies obstruct the search to begin with.

In running test searches, I also uncovered troubling instances of suppressed results. I recently saw the Oscar-winning documentary No Other Land, co-directed by Palestinian and Israeli filmmakers. I decided to ask Research Assistant about this film. Searches for most of the directors generated the same no results error messages. A search for one of the co-directors generated partial and inaccurate results. Research Assistant made several guesses as to what the film No Other Land was about and why it was important. Could this be an example of biased third-party policies preventing searches about this Palestinian film and filmmakers? Is it also possible that Research Assistant is accurately reflecting the lack of materials related to No Other Land available in Ex Libris’ entire Central Discovery Index and widespread censorship of this film (and more broadly Palestinian people, culture, and experiences) in the United States? Are there truly no academic sources talking about this film and these filmmakers, one of whom was recently attacked outside his own home by Israeli settlers? More research is needed to clarify why “controversial” searches return error messages, inaccurate results, or partial results. The current state of Research Assistant partial results and error notices for “controversial” topics warrants more inquiry.

I know with near certainty that if a student user at my workplace received an error message when searching for “Tulsa race massacre,” they would switch their topic and probably not tell me or their instructor about the error message. To my users, “we couldn’t generate an answer for your question” translates to “your topic is not worthy of pursuing—change it.”

Uncritically adopting AI tools in discovery systems will perpetuate, if not exacerbate, existing biases and suppression of minoritized people. Try this safer topic. Try this approved topic. Try this unobstructed topic.

Libraries, museums, and information sciences are a frontline in resisting the federal administration’s targeting of minoritized people as well as their whitewashing, disinforming, censoring, and defunding tactics. In the age of AI, our access to information is not spared from these attacks. We must uncover answers to the questions Safiya Noble, Joy Buolamwini, and others elicit: Who and what is safeguarded through AI tools in discovery searches? Whose safety and comfort are prioritized? At whose expense?

…so what exactly is going on between publishers and the NIH?

We’ve reached an important historical inflection point in the funder/author/publisher triangular relationship.

Background

If you work at a university, you’re likely aware of the current presidential administration’s dramatic mandates to federal agencies that fund research, most infamously requiring them to cancel thousands of active grants related to diversity, equity, and inclusion, and attempting to significantly decrease indirect costs paid to universities for overseeing federal grant compliance.

Interestingly, however, the current administration and Director of the Office of Science and Technology Michael Kratsios have given no indication that they will backtrack mandates laid out by the previous administration’s Nelson Memo, which advised federal funders that grant recipients should make their funded research outputs publicly available without an embargo period (among other things). The goal of the Nelson Memo was to “…promote the rapid sharing of federally funded research data with appropriate protections and accountability measures [that] will allow for greater validity of research results and more equitable access to data resources,” along with the rationale that “There should be no delay between taxpayers and the returns on their investments in research.” This memo was widely seen as a major win in the open knowledge community, and also apparently meets the approval of the current presidential administration and their appointees (more on that below).

For those unfamiliar, publishers of scholarly journals often require authors to agree to an embargo period, which means no version of their article may be made publicly available on any other platform for a certain amount of time (usually 6 to 24 months depending on the journal). This allows publishers to profit from subscription-only access to new research during its period of highest readership. After the embargo period, publishers have generally allowed authors to make available the “author’s accepted version” of the article (the version after peer review but before publisher copyediting and typesetting) in a funder’s repository, institutional repository, or disciplinary repository. The practice of sharing items openly in this way is often referred to as “self-archiving” or “green open access.” It is also very common for journals to offer a pay-for-open model, through which authors can pay an article processing charge (APC) to apply a Creative Commons license to their article instead of transferring all rights to the publisher. This is often referred to as “gold open access.” Federal funders have typically allowed authors to use grant funds for APCs.

The National Institute of Health (NIH) has long required grant recipients to deposit copies of grant-funded articles in their openly accessible PubMed Central repository, but until this year, the NIH allowed publishers to retain sole distribution rights for a 12-month embargo period if authors chose not to pay for open access. In turn, many publishers have offered authors the service of depositing the appropriate article version in PubMed Central on their behalf after the embargo period ends.

Showdown

Last year, in compliance with the Nelson Memo, the NIH announced an updated Public Access Policy, originally slated to take effect on December 31, 2025. On April 30, newly appointed NIH Director Jay Bhattacharya announced a 6-month acceleration of the policy’s effective date, to begin on July 1. The new policy “…requires Author Accepted Manuscripts accepted for publication in a journal, on or after July 1, 2025, to be submitted to PubMed Central upon acceptance for publication, for public availability without embargo upon the Official Date of Publication.” Unsurprisingly, it seems several of the largest publishers are not happy about this new public access policy.

As an early career librarian trying to parse this complicated, evolving situation, I’ve relied heavily on my professional communities. Conversations on the University Information & Policy Officers listserv have been really helpful for my understanding. Jeremy York, Assistant Director of the Copyright Office at University of Michigan Library, kindly granted me permission to quote his summary of publisher reactions.

“Publishers are responding to the new policy in different ways. Some publishers, like Wiley and Springer Nature will no longer deposit works in PubMed Central on behalf of authors unless authors pay an Article Processing Charge (APC) to make the work available under an open license. They also are not allowing deposit by the author of the author accepted manuscript (AAM) in PubMed Central and see the pursuit of this ‘green open access’ route by authors as a violation of their publishing agreement.

The American Chemical Society (ACS), on the other hand, has implemented an Author Development Charge (ADC). If authors do not pay the APC to have their work published with an open license, they can still pay the ADC to be able to deposit the author accepted manuscript (AAM) in PubMed Central to comply with NIH’s policy.[1] If authors deposit the AAM in PubMed Central without paying the ADC, ACS sees this as a violation of the publishing agreement.”

The complicating factor here is that the NIH’s claims over funded research outputs predate any author’s agreement with a publisher. The NIH is simply exercising a right that all federal funders have held for a long time, called the Federal Purpose License. In the NIH’s case, their version of federal purpose license states grant recipients “…may copyright any work that is subject to copyright and was developed, or for which ownership was acquired, under a Federal award. The HHS [Department of Health and Human Services] awarding agency reserves a royalty-free, nonexclusive and irrevocable right to reproduce, publish, or otherwise use the work for Federal purposes, and to authorize others to do so.”

In response, certain publishers are advising authors that the only way to comply with both a) the NIH public access policy and b) their journals’ copyright agreements is to pay some sort of fee. They are signaling an unwillingness to publish paywalled articles that will be made publicly accessible on another platform; if a preexisting claim to an article exists, then publishers may simply choose not to publish that article despite acceptance by an editorial board, unless the author pays a fee.

Yet, the purpose behind federal funders’ public access policies is not to force authors to pay for open access, as evidenced by the NIH’s announcement on July 8, which revealed plans to cap the amount of NIH-funded dollars that authors can spend to make articles open access (or in other words, cap the amount authors can spend on APCs) beginning in 2026.[2] Clearly, the NIH does not want to pay exorbitantly for articles it already holds rights over via the federal purpose license.

Breakdown

This is an important historical inflection point in the funder/author/publisher triangular relationship, in part because the NIH awarded over $26 billion across 41,304 research project grants in 2024, up from $22 billion in 2019 (although those numbers may deteriorate going forward given the current presidential administration’s proposed budget). Needless to say, federal funding is integral to the United States’ research infrastructure, meaning federal funders possess a lot of leverage to shift that infrastructure. What happens over the coming months as other federal funders’ public access policies take effect may significantly alter available open access publishing pathways for scholarly authors and influence subscription negotiations for libraries.

For interpretations of what this ongoing situation means for authors and advice about how to navigate the publication process as a grant recipient, I recommend Authors Alliance, who put together this excellent FAQ page [and update here] on the current “contradictory policy environment,” and are co-hosting a webcast on August 5, 2025 focused on new federal funder sharing requirements.

For academic libraries, one likely outcome in the immediate future is increased traffic on existing read-and-publish deals and increased pressure to engage in new and/or bigger read-and-publish deals from research faculty.[3] However, many academic libraries receive some portion of indirect costs paid by federal funders to universities for grant oversight, so any budget squeeze due to the White House’s proposed cap on indirect costs would make it difficult for those libraries to invest in new or more expensive read-and-publish deals.

On the other hand, if both publishers and federal funders were to hold firm in their current standoff, public access policies and publication fee caps could gradually redirect authors away from pay-for-open publishing models and toward outlets that allow for self-archiving or use a free-to-read, free-to-publish model (sometimes referred to as “diamond open access”). Big for-profit publishers are relying on their journals’ prestige and importance in tenure portfolios to withstand these shifting sands, but if the cost of publishing in their journals remains unaffordable for authors, readership and tenure committees could begin to elevate free or affordable open access journals in their place. Many such alternative outlets exist; for journals with zero-embargo deposits, you can check JISC open policy finder, and diamond open access journals are easily found on the Directory of Open Access Journals by selecting the “without fees” filter. It is worth noting that, among the big publishers, Sage stands apart by providing a relatively generous zero-embargo self-archiving policy.

A world in which libraries regain some financial relief from big publishers, and in which more people can benefit from openly accessible research findings, is one we should hope for. The actual outcome of this standoff is impossible to predict but is certainly worth monitoring for academic librarians.

[1] Several smart folks, including Authors Alliance, have doubts about whether authors will be allowed to use NIH grant funding to pay for an ADC. From Jeremy York’s summary: “The NIH guidance states that ‘Costs for publishing services that are charged differentially because an Author Accepted Manuscript is subject to the NIH Public Access Policy or the work is the result of NIH funding are unallowable because charges must be levied impartially on all items published by the journal, whether or not under a federal award’ (GPS 7.9.1).”

[2] The announcement did not indicate an effective date or exact cap amount.

[3] For those unfamiliar, read-and-publish deals (sometimes referred to as “transformative agreements”) between libraries and publishers provide a mechanism for an institutions’ authors to publish through an open access pathway without themselves paying APCs. Instead, the institution can draw from a pool of funds or credits to cover the APC on behalf of the author. This open access mechanism is part of the library’s subscription to the publisher’s journal portfolio. Aside from libraries shouldering massive subscription costs, the main drawback of read-and-publish deals are that libraries are often subscribing to journal access as opposed to purchasing perpetual access for that year’s journal issues. This move toward conditional access and away from ownership creates an increasingly difficult situation for libraries to extract themselves from, lest they lose access to journal backlogs from the years they subscribed instead of purchased.

Rethinking Wikipedia from a Library Perspective

Editor’s note: We welcome a guest blog post from Nathan Sonnenschein, User Services & Experience Librarian at Montana State University Billings.

During my first year as an undergraduate, my community college’s library handed out folders bearing a catchphrase: “Faster than Google. Better than Wikipedia. Your college library.” This gave me the impression that academic libraries saw Wikipedia as a usurper on their terrain—an upstart competitor of the digital age. This struck me as out-of-touch. I had to start my research somewhere, and Wikipedia was an easy, familiar way to learn the essentials about a topic before searching the library for books or articles. Furthermore, checking Wikipedia’s citations helped me find crucial sources I might have otherwise missed. I never would have cited Wikipedia directly in my papers, but then again, I would never have cited any encyclopedia.

In the decade since, I have been delighted to see many library professionals develop more nuanced approaches to Wikipedia. Past ACRLog posts reveal a few of these perspectives. Wikipedia is a handy partner for brainstorming, a clever means for disseminating new research, and a great case study for teaching info-lit students about systemic biases or the facets of the ACRL Framework. Many libraries over the years have even hosted edit-a-thons and other initiatives to promote active, critical engagement with the online information landscape.

Still, the belief that Wikipedia has no place in research or scholarship persists. Many of my students’ zealous high school teachers taught them to never use Wikipedia, ever. To be fair, Wikipedia is far from perfect. Its mechanisms can and do enable vandalism (though editors often revert vandalism within a few hours, or even a few minutes). Readers must watch out for uncited statements or promotional language. For new editors—Wikipedia’s term for its contributors—the site’s policies and style guidance can feel frustratingly arcane and difficult. Wikipedia’s editor base skews heavily white and male, which introduces the potential for unconscious bias, and can create an alienating environment for editors who are women and/or people of color. These are real issues, but they do not mean Wikipedia has no valid place in the information ecosystem. Any information source has its strengths and weaknesses, and we must approach all sources with a critical lens.

Optimistically, I hope the library field has reached a place where most of us understand how great Wikipedia is for certain uses. However, we need to go further. In our new world of generative AI and volatile American politics, our fields must approach Wikipedia as not just a tool, but as an ally. Let me explain.

A shared mission

Generally, the digital revolution in the library sector has been intertwined with the trends of corporate growth and consolidation. Academic libraries’ collection development funds funnel almost entirely to massive publishers and “information cartels.” These companies have a vested interest in keeping information behind paywalls, and in protecting their own rights. This pursuit of profit is actively at odds with our libraries’ economic realities—with flatlined or diminishing budgets, serving our communities’ information needs is increasingly challenging. (Quite frankly, it feels unethical to subsidize these corporations’ profit margins with the limited funds my library receives—especially when that funding comes from student fees.)

In contrast, Wikipedia is run by the Wikimedia Foundation, a nonprofit organization that seeks to “help everyone share in the sum of all knowledge.” (This is a big reason why Wikipedia has not suffered from the enshittification plaguing Google Search and other platforms.) The library field and the Wikimedia Foundation are both rooted in the belief that individuals benefit from information access. We are not competing, but working toward the same vision. As library professionals, we are not seeking to enrich our institutions by boosting profits. We are seeking sustainable mechanisms for keeping our communities connected to information.

The political climate

Just as libraries and Wikipedia have similar goals, we are facing similar challenges. On April 24, 2025, the Wikimedia Foundation received a letter from the U.S. Department of Justice claiming the organization may have violated nonprofit laws. Why? The letter levels several accusations: spreading propaganda, rewriting the details of events in American history, and failing to provide “neutral” educational content—content that allegedly benefits foreign powers. The inference is that Wikipedia must engage readers with a point of view that directly favors the United States, and should not even have “foreign nationals” on its board, despite Wikipedia’s international scope.

One passage in particular sticks out to me: “Masking propaganda that influences public opinion under the guise of providing informational material is antithetical to Wikimedia’s ‘educational’ mission.” If you swap “Wikimedia” for “libraries,” this statement rings similarly to the accusations unrelentingly leveled against libraries over the past few years. Whether it be the supposed spread of anti-American narratives, the “indoctrination” of children via books with queer themes, or the dissemination of scholarly materials on Critical Race Theory, libraries are challenged for providing information that does not align with the agenda of certain powerful political factions. While the DOJ’s accusations against Wikipedia differ in form from the censorship pressures faced by libraries, the underlying logic—targeting institutions for disseminating perspectives that do not align with a certain political agenda—is strikingly similar. The result? Even as federal library grants are revoked, the Wikimedia Foundation has its nonprofit status threatened.

The rise of AI

The information landscape is quickly shifting as generative AI tools, like ChatGPT and Google Gemini, are revolutionizing information discovery. For libraries, this just means adapting to yet another wave of technological innovation. However, because Wikipedia is relatively young, its resilience is being tested. There has already been some doom and gloom about Wikipedia’s continued existence in a world where AI can conveniently provide encyclopedic information. Yet, many AI platforms are heavily dependent on information sourced from Wikipedia (and from libraries’ digital collections, for that matter). This creates a new paradigm: an information landscape where AI-generated information proliferates, but traditional, human-generated sources—primary, secondary, and tertiary alike—take on a new dimension of reliability and authority. The latter cannot exist without the former.

In other words, with so much machine-generated content on the internet, Wikipedia’s content takes on new value. The Wikimedia Foundation recently reaffirmed its commitment to “putting humans first.” On the editing side, this means relegating AI to technical tasks while continuing to rely on real people for knowledge curation. For the everyday reader, Wikipedia remains committed to creating for the sake of human learning, not for AI ingestion. Meanwhile, libraries continue to exist for the purpose of serving people, prioritizing access, privacy, and reliability. To say that another way, Wikipedia and libraries put people first—not for the sake of profit, but for the betterment of society.

What does this mean?

If I were a first-year college student today, I hope that the librarians would not tell me to ignore freely available online research tools in favor of books or databases. Rather, I hope they would nudge me to navigate all information sources critically and thoughtfully—including Wikipedia.

Why? Wikipedia is not perfect, but it is one of the last vestiges of an earlier, idealized vision of the World Wide Web: a place where anyone can sail the wide-open seas of human knowledge. We cannot take it for granted. Even as the information paywalls in the scholarly world continue to grow higher, Wikipedia remains an incredible resource for discovering information written by humans, for humans. In an information ecosystem dominated by corporate interests, information paywalls, and tech oligarchs, libraries must stand with Wikipedia as catalysts for a more equitable future.

Open Letter to CRL from the academic wing of #CripLib

Editor’s note: We welcome a guest blog post from the academic librarians of #CripLib, which focuses on the intersection of disability and library work.

Note: The focus is on our concerns about the quality of the article and the editorial process that led to it being published as research – NOT on the author as a person.

In Brief

The authors of this letter are writing in regard to “Respecting Privacy of Thought in DEI Training,” an article published in the May 2025 issue of College & Research Libraries (CRL).

How did such an intellectually dishonest, poorly argued editorial get published in College & Research Libraries, a highly-regarded library journal? We will identify some of the most obvious problems within the text–problems that any competent editorial team should have addressed before publication. As part of this, we also ask: why did the editorial team not support this author in refining her argument before publishing it? Or alternatively, why did the editorial team not reject this article?

Full version

Reason for writing this letter

Despite approaching the article with curiosity, we quickly identified several issues which caused us considerable concern:

The author neglects to engage with other scholarly critiques of DEI in library literature.
The political bias is hidden.
The author appears to misunderstand or shallowly understand several of the cited sources.
Citations are spotty – while some arguments provide adequate citation, other arguments lack evidence.

We are members of what are often referred to as marginalized populations–we all have different disabilities, and many of us are also LGBTQ+ and/or BIPOC, etc. We have been dismayed by the recent dismantling of DEI initiatives, which is operating hand-in-hand with the dismantling of our civil rights. We are also very open to critiques of DEI programs – we and people like us have often been at the forefront of critiques about DEI programs in libraries, expressing concerns about tokenism, lack of follow-through, ineffective “one shot” style training models, prioritizing white feelings over the material conditions of Black and Indigenous people or people of color, to name a few (see References).

However, none of the many articles written by librarians articulating critiques of DEI initiatives are engaged in this article. Instead, the author focuses on engaging with a frankly odd mix of Wikipedia articles, shallowly analyzed law review articles and political theory books, and right-wing blog posts and media. We recognize that several of these right-wing media sources identify as libertarian and that Anglo-American style libertarianism often refers to itself as non-partisan, but that doesn’t mean it lacks an ideology (tone marker: neutral valence, see Roberts, 2025). Meanwhile, accusations of library DEI initiatives being “politically partisan” are made a few times, without evidence. (Note: no definition is provided for “politically partisan,” so we are working on the assumption that it refers to allegiance to a political party.)

Frankly, we are concerned by the use of recent far-right talking points to argue against DEI initiatives in libraries without an honest acknowledgement of the political stance the author is taking. To be clear, we are not objecting to the author having a political stance, we are objecting to hiding it. Among those of us drafting this letter, we do not share the same political beliefs about everything, we just share the expectation of honesty and academic integrity, so we can actually work through those disagreements in a productive and respectful way.

Even with the current work being done to dismantle DEI, civil rights, and higher education at the direction of one political party, we would also be among the first to question political partisanship in libraries. That is because none of the dominant political parties in English-speaking North America reliably support us (see Anyangwe and Okoh, 2023; Dhenin, 2025 for two recent examples of bipartisan legislation that causes active harm to disabled and other marginalized peoples).

However, whatever our feelings are about the author’s conclusions, what we’re really dismayed by is its failure to honestly engage with legitimate critiques of DEI initiatives in libraries. There are some elements of the piece that we may have agreed on, if they’d been presented with more academic rigor.

That leads us to wonder: How did such an intellectually dishonest, poorly argued editorial get published in College & Research Libraries, a highly-regarded library journal? We will identify some of the most obvious problems within the text – problems that any competent editorial team should have addressed before publication. As part of this, we also ask: why did the editorial team not support this author in refining her argument before publishing it? Or alternatively, why did the editorial team not reject this article?

Problems identified within the text

Because the errors in argumentation and the questionable citations are so rampant within this article, we are puzzled by its publication.

The article includes a lot of references. We wonder if that was required for it to be called research rather than a letter to the editor? In any case, it has resulted in a long list of cherry-picked, manipulatively quoted sources that support what the author wants to say. This piece does not demonstrate any honest engagement in ongoing conversations in the field or even the arguments made in the cited works.

Similarly, the author demonstrates limited understanding of the academic texts referenced. Some of the references are to right-wing blog posts. Note that we’re not condemning the use of blog posts or popular press books across the board, what we’re critical about here is the relative lack of academic citations and the lack of disclosure about the political biases of the rest of the sources that were used. When talking about DEI initiatives at a time when the US federal government is actively dismantling them, while simultaneously engaging in anti-Black, anti-immigrant, openly eugenicist, and anti-LGBTQ+ initiatives, this matters. The author does not clearly identify the partisan nature of many of her non-academic sources–partisan as in partial to an approach that is generally viewed in the US as right-leaning, is often critical of social justice causes, and in some cases openly advocates for the right to discriminate (e.g., Levy, 2016).

Here are a few sections that we have selected to highlight, just for demonstration. This is not an exhaustive exploration of all the weaknesses of this article.

Does not engage with other scholarly critiques of DEI in library literature

Tracking attendance at trainings is one new data point, especially where participation is required or assessed in performance reviews.¹¹⁰ DEI action plans, rubrics, “temperature checks,” climate surveys, equity audits, assessments of impacts, and so on, in aggregate contribute to shifting the library’s culture away from the academic and toward the bureaucratic.¹¹¹ The message they send to the library workers is that their own independent thoughts and autonomy in their jobs is secondary to data-collection tasks.

One example of decentralized, bottom-up control is known in the organizational management literature as “concertive control.” Under concertive control, organizations employ cross-functional and self-organizing teams.¹¹² Library DEI trainings, while potentially administratively sponsored, are often delegated to a team that is responsible for their content and that may also conduct the trainings. Library administrators will naturally delegate DEI training to those who are motivated and well-versed in the ideas. These trainings, and the teams that lead them, are a mechanism for non-hierarchical, concertive control over employees. Teams that do not have authority over library policy may be given wide latitude to develop trainings, especially when they can be based on profession-wide guidelines…. (p. 442)

The author has not engaged with any of the many and varied library articles about managerialism, workplace climate, library assessment, etc. The implication of this section appears to be that library administrators should not delegate DEI trainings, in fact should not be using concertive control to delegate anything, because to do so means they lose all responsibility and control. If this is indeed what readers are supposed to take from this section, we would have liked a clearer articulation of — everything, frankly. If the author is not arguing that library administrators should not delegate work, that raises even more questions. Also, again, if the author had wanted to build on arguments made previously by other librarians critiquing various management tendencies, we think that would have enhanced and clarified her argument.

Does not disclose political bias

While this sort of statement is not required, it matters for this discussion because we are talking about issues where people hold a wide range of conflicting values and beliefs, and the argument is presented as if it is coming from some sort of objective place.

Pretending that you can talk about race and racism from a place absent ideology is weird. But if you take the perspective of someone who doesn’t want to think or learn about racism, it starts making more sense.

The author praises the 2012 Diversity Standards for demonstrating “non-ideological language” while criticising the narrower focus of the 2022 Cultural Proficiencies for Racial Equity (p. 439). Shortly afterwards she claims that “…. imperialism and colonialism, mass incarceration, disparities in access to healthcare, the achievement gap in education, and disparate outcomes for “BIPOC populations.”⁹⁵ None of these topics has more than an indirect link to library work.” (p. 440) There is plenty of evidence out there identifying how each of these points connect to each other (see References). A good place to start looking is within the Cultural Proficiencies. This section provides one of many examples of the author pushing back specifically on training that requires attendees to learn about racism. We can’t help reach the conclusion that this is not an absence of ideology, this is a refusal to acknowledge one’s own.

Misunderstand or shallowly understood source material

Much of the article centers on the idea of privacy of thought, the author’s definition of which is:

Privacy of thought is an instinctively understood concept because everyone experiences it every day. Most would consider being forced to share private thoughts an invasion of privacy. In this way, the idea that privacy of thought should be considered a human right feels intuitively correct. How this intuition is reflected in society via laws and culture, however, is inseparable from the influence of Enlightenment ideas about how we understand individual liberty…. (p. 432)

The author then goes on to use various references to argue that freedom of thought is a universal value, followed by a reference to cognitive liberty. We assume that the author has done the internal work of synthesizing these concepts, and that’s why they’re presented in conflated form. However, that work is not replicated on page, so we as readers are confused about why these different terms are treated as interchangeable.

Also, nowhere in this section is an acknowledgment that the Farhany discussion about cognitive liberty is specifically in reference to tort law, the legal doctrine of avoidable consequences, and emotional distress injuries. This is a really different context than workplace DEI training. We recognize that a link could be made here, but the author did not actually make it – for us this raises the question about whether the author read and understood the source material. This is a rich enough area for investigation that she could have made a whole paper focused on this idea. Instead, the paper immediately moves on to mostly unsubstantiated claims about DEI trainings (almost exclusively focused on critiquing anti-racism trainings), with yet more under-explained theoretical frames added to the discussion. And as we observed earlier, there are articles published by librarians on these issues, none of which are engaged here.

When talking about a self-assessment conducted as part of a racial equity training, the author notes that:

“Even if your employer does not see your responses, these ninety-six questions are presumptuous and invasive. Asking about a person’s thoughts in this way violates privacy of thought…” (p. 435).

Is learning something new or considering new approaches an invasive violation of privacy? It seems absurd on its face. Is having to think about challenging things – not having to publicly disclose your thoughts, just having to think about them – a violation of privacy? What are the boundaries of this argument – is it only when people are expected to work towards ending racism that it becomes a violation? This is a really strange argument that none of us understand. There may be a sensible rationale behind this argument, but it needed to be explained within the article before publication.

Also the source referenced in the above excerpt was a right-wing/libertarian blog post, rather than a direct reference to the assessment tool itself or a first-hand account. It’s unclear whether the library worker quoted within that blog post had actually taken the assessment. It is possible that this blog post was writing about one employee who just didn’t like how the Inclusive Excellence frameworks were implemented at UVA. In other words, the author’s argument here relies on unclear secondhand information provided by one person.

This section included a side point that “(Suggesting a redirection back to race and racism here implies that other marginalized groups are not appropriate DEI topics.)” – but all the citations in this section are to anti-racism trainings. Asking people to stay on topic seems like a reasonable request for a training session. Even if one wants to make this point about DEI trainings in general, which some have (see References), to once again make the focus of the argument a pushback against anti-racism training is noticeable. Again, why did the editorial team not catch this? Spotting this problem does not even require anyone to dig into citations, it’s self-evident just from reading the text itself.

Next steps

We are asking for a public statement from the CRL board articulating how they will rectify this situation and make improvements to prevent a recurrence.

Problems identified:

The author neglects to engage with other scholarly critiques of DEI in library literature. There have been innumerable articles, books, and presentations that talk about DEI in relation to the concerns brought up in this article (e.g., assessment of DEI programs, managerialism, the tensions between intellectual freedom and social justice) but none of that body of research was used here.
Instead the author uses sources that broadly fall under the spectrum of political theory that, as demonstrated in the text, appear to be misunderstood or shallowly understood.
The citations provided are plentiful but poor in quality, including more wikipedia articles than library research articles. While some arguments provide adequate citation, other arguments lack evidence.
We all have biases. Pretending that we or the systems we work with don’t operate in biased ways does not erase those biases, it just prevents those biases from being acknowledged or addressed. This article talks about DEI in libraries while pretending at objectivity in its framing, but the arguments presented and the citations referenced demonstrate a clearly biased point of view that is never acknowledged.

Because of the degree and pervasiveness of the problems we identified as well as the potential negative impacts on the field if this article is taken at face value, we are asking CRL to retract this article. We would also like an explanation: how did this make it through editorial review? Is there a peer review process at CRL? Are you having subject experts conduct peer review? Are you outsourcing peer or editorial review to artificial intelligence?

References

Anyangwe, D. and Okoh, C. (2023). The Bipartisan Safer Communities Act: A dangerous new chapter in the war on Black youth. The Center for Law and Social Policy. https://www.clasp.org/publications/report/brief/the-bipartisan-safer-communities-act-a-dangerous-new-chapter-in-the-war-on-black-youth/

Courtland, S.D., Gaus, G. and Schmidtz, D. (2022). Liberalism. In Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Spring 2022 Edition). Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/liberalism/

Dhenin, M. (2025, May 9). Escalating anti-homeless policies fall hardest on disabled people in the US. Truthout. https://truthout.org/articles/escalating-anti-homeless-policies-fall-hardest-on-disabled-people-in-the-us/

Joint ALA/ARL Building Cultural Proficiencies for Racial Equity Framework Task Force. (2022). Cultural proficiencies for racial equity: A framework [White paper]. ALA. https://www.ala.org/pla/initiatives/edi/racialequityframework

Levy, R.A. (2016). Libertarianism and the right to discriminate. CATO Institute Policy Report. https://www.cato.org/policy-report/march/april-2016/libertarianism-right-discriminate#

Roberts, W.C. (2025). Ideology. In Zalta & Nodelman (Eds.), The Stanford Encyclopedia of Philosophy (Spring 2025 Edition). https://plato.stanford.edu/entries/ideology/#ValIde

(see References) note: We had originally referenced the work of respected librarian-scholars who’ve published research in this area. However, because we value them so much, and because we recognize the risk of speaking out publicly like this, we decided not to publish their names. Those new to this sort of work may not know the history of librarians doing DEI work being targeted by harassment campaigns and death threats – especially women of color – but those of us who have been around for a few decades don’t want to see that happen again.

Reconsidering research impact

I have a complicated relationship with bibliometrics (i.e. “research impact metrics”).

The silhouette of a person looking at a huge swirling modern art piece on a digital screen. — Photo by Gunnar Ridderström on Unsplash

I have a complicated relationship with bibliometrics (i.e. “research impact metrics”), as does the field of academic librarianship and academia itself, if we’re honest. On one hand, there is something undeniably fascinating about how statistical formulas can help us understand “scholarship as conversation” at a macro-level. As a field, bibliometrics has been heavily influenced by information scientists. Eugene Garfield, often cited as a founder of modern bibliometrics and cocreator of Journal Citation Reports, held a degree in Library Science from Columbia University, for example, and created the infamous journal impact factor not to assess the quality of a researcher’s CV, but to help librarians make decisions about which journal subscriptions to purchase. This relationship with librarianship makes me feel some level of responsibility to/for the field of bibliometrics even if I’m not a bibliometric scholar per se. Librarians’ collection development responsibilities have traditionally positioned us as arbiters and assessors of information quality, for better or worse, which tracks with the project of research assessment.

On the other hand, several high-profile bibliometric formulas have been coopted by corporate entities to serve for-profit purposes and have been misapplied in the academe as markers of excellence for individual researchers. In the 2010s, research assessment experts denounced the use of Clarivate’s journal impact factor and overreliance on quantitative citation-based evaluation of research “impact” through statements like the San Francisco Declaration on Research Assessment (now operating as DORA) and the Leiden Manifesto for research metrics. I share concerns cited in those statements about whether folks who use these numbers in retention, promotion, and tenure (RPT) processes can accurately explain the insight each formula is supposed to convey. Journal impact factor and Elsevier’s “CiteScore” are not generic proxies for journal quality. They aggregate citation rates over a (relatively short) period. In a vacuum, that seems like potentially useful information to consider as one part of a comprehensive assessment of a journal’s and/or a researcher’s portfolio. As a key determiner of whether a journal or its authors offer any value to the scholarly conversation, it becomes dangerous.

It is also well established that these popular metrics don’t map well across academic disciplines because different disciplines have significantly different publishing rates and citation practices, whereas metrics that normalize citation attention between disciplines like field-weighted citation impact are not commonly used in RPT procedures (from my understanding). Consequently, the perceived value of scholarship in the arts, humanities, and social sciences is diminished when their citation numbers are compared to STEM fields. In the fields that bibliometrics favor, their outsized influence in RPT processes forces faculty to “study for the test,” thereby limiting the list of suitable publication venues to a select, well-established few (often owned by one of the big five publishers). From a simple assessment standpoint, Goodhart’s law reminds us “when a measure becomes a target, it ceases to be a good measure.”

Although Goodhart’s law is certainly applicable to the current use of bibliometrics in academia, I think the problem requires us to zoom out a bit. The core issue leading to the misapplication of bibliometrics is academia’s attempt to assess a concept as sweeping and nebulous as “impact” using a few numbers that address one specific type of impact—contribution to the scholarly conversation. Furthering the progress of knowledge is a valid and core reason why many scholars do research, but it certainly isn’t the only reason why people do research, and it’s not the only value that research provides to the world, so it shouldn’t be the only aspect of research impact we evaluate and thereby ascribe value to.

Of course, making any kind of structural change in academia is easier said than done for many reasons, including the entrenchment of bibliometric-based evaluation in the publishing industry and in academic tenure processes. But, there are folks exploring alternatives: in 2023, the University of Illinois Urbana-Champaign introduced a Public Engagement Research Option (PERO) for Promotion and Tenure, designed for faculty who “address a societal problem to contribute to the public good; produce traditional and non-traditional outputs; collaborate with communities or organizations; [and] engage in a mutually beneficial exchange of knowledge and resources with community partners.” The HumetricsHSS Humane Metrics Initiative provides workshops and toolkit resources for academic departments to consider how they can embed their stated values in their evaluation of scholarly life, reminding us, “if we don’t measure what we value, we will only value what we can measure.” Just this month, DORA published their Practical Guide to Implementing Responsible Research Assessment at Research Performing Organizations.

I heard about that DORA resource while attending a new (to me) conference in April. The Center for Advancing Research Impact in Society hosts an annual summit around “broader impacts”—the National Science Foundation’s terminology for research’s “potential to benefit society and contribute to the achievement of specific, desired societal outcomes.” I’d been debating which research impact professional development opportunities I should explore, given that “research impact” is included in my job title. This one intrigued me because of its necessarily broad conception of impact, and I came away energized by the creative research development work happening around the world. Research development offices are logically very interested in understanding and tracking the impact of research conducted at their institutions, and they clearly spend a lot of time thinking about how to do research assessment well, and how to encourage researchers to embed the public good in their plans for research outputs. Two favorite sessions included the Association of Public & Land-grant Universities’ walk-through of “Modernizing Scholarship for the Public Good: An Action Framework for Public Research Universities,” and the Research Development team at University of Texas at Austin’s presentation on their Research-to-Policy workshops that teach faculty how to make their work accessible to legislators.

Although librarians are welcomed at the ARIS Summit, I didn’t meet any other library folks wandering between conference rooms. I mention this not to admonish anyone or toot my own horn, but rather to raise the question: why shouldn’t we participate at these types of conferences, contributing to and learning from conversations about research impact/assessment? I would argue the prevailing (mis)conception of research impact is an information literacy issue. As a librarian, I feel a responsibility to remind scholars of the ways “authority is constructed and contextual” and how our interpretations of information (like bibliometrics) influence the way those specific pieces of information come to have value. These issues are especially relevant to those of us in the scholarly communication subfield, as we play a particular role in helping scholars decide where to communicate and share their work. I now see the “open knowledge” and “research impact” components of my job title as much more intertwined that I originally thought, but it required me to conceptualize a holistic vision of research impact that incorporates the societal contribution we believe many types of scholarship can make. Academic libraries are increasingly involved at every stage of the knowledge lifecycle, and as such, I think research assessment/evaluation is an area where librarians could reconceptualize our approach, training, responsibilities, and campus partnerships.