Looking to Transition to Plan B?

Back in the dark past of 2018, Plan S launched with the objective of destroying the business models of academic commercial publishers and making publicly funded research open and 'free'. The Plan S Open Access model (OA) was long anticipated but still shocking to the industry when it finally came. Like the grunts on Omaha beach, journal publishers could be forgiven for thinking "This is it. We may not come out of this one".  The Plan S vision foretold a world where research funded by Coalition S members would be published fully OA without embargo periods. They even specified a date(s) when this 'transition' had to be completed. Plan S promised to end the subscription-based model with the scholarly publishing community confronting a fundamentally changed business environment.

Plans go awry once they start shooting at you.

Unquestionably, some objectives of Plan S have been met. There does appear to be significantly more OA publishing is some scholarly fields - particularly physical sciences. Are these collections any easier to access? That is debatable. Have the largest commercial publishers successfully adapted their business models, continued to consolidate and entrench their positions? Undoubtedly, yes. Are society publishers under greater financial pressure?  Harder to answer, but almost certainly their financial options are less favorable than the pre-Plan S years: Fewer large dollar consolidations with commercial publishers, less ability to leave deals, less money to spend on society objectives.

Yet given this back ground, the recent announcement by PlanS still comes as a surprise and signals something like a retreat from the founding principles of OA as a transformational construct. The new directive allows for concessions which weaken the original premise: Hybrid journals are now allowed which presents a significant concession to both publishers, funders and researchers who have been unwilling to carry the full costs of OA publishing. Notably, libraries also face both real and administrative costs in an OA workflow which is rarely fully acknowledged. We now seem to have a less dogmatic and inflexible approach but one which draws into question the very point of the OA effort since 2018.

What does it all mean?

In the short team, probably not a lot. Although in some publishing board rooms there may be some difficult questions raised over the next few months about the business strategy. In a hybrid world, large commercial publishers will extend their size and market clout and Article Processing Charge (APC) and traditional subscription models will carry on uninhibited. Arguably, looking forward we may be in a worse situation vis-a-vis affordability and access than before Plan S started. The gap between underdeveloped and underfunded may well be wider: Since 2018 we have had further business consolidations and the subscription + APC pricing models are now entrenched. Did we really believe commercial publishers would roll over?  Elsevier's operating margin is estimated to larger now than it was five years ago.

Perhaps this Plan S announcement will be viewed as a tactical adjustment, an insignificant bump in the road to transformative OA but it really doesn't seem like it. What began as a Teutonic dictate to the industry at large, now looks more like a research article which hasn't made it out of peer review. It probably will not be published.

Black Box, White Collar: How AI Learned to Reject You in Milliseconds

A recent research article published in APA''s Journal Consulting Psychology Journal titled, Some Ethical and Legal Issues in Using Artificial Intelligence in Personnel Selection by Dr. Olaf Ringelband and Dr. Christian Warneke looked at the use of AI in the hiring process. The article can be found on psycnet.apa.org

If you have experienced AI in the sandbox which is LinkedIn you might have a predisposed view - that is negative, on the use of AI in the hiring process. But more and more companies are using AI not only to match and screen but also to interview candidates for important roles. For my part, thanks to LinkedIn's matching system, I now get daily notifications about exciting opportunities that require 10 years of experience in technologies that were invented 3 years ago, which really makes me feel like I'm nailing this whole career thing.

The Problem with Algorithmic Hiring 
Ringelband and Warnke remind us that artificial intelligence is rapidly reshaping the landscape of personnel selection. What began as a promising tool for streamlining recruitment has evolved into a complex ethical and legal minefield. AI systems, trained on historical data, now influence hiring and promotion decisions across industries. Yet the opacity of these systems—particularly those based on deep learning—raises troubling questions. Their “black box” nature makes it difficult to discern how decisions are made, and whether they are fair. The risk is that AI may not only replicate existing biases but institutionalize them, all under the guise of efficiency.

The authors of this article set up the discussion in the following way:

In the present article, we aim to translate research into practice by highlighting the points of contact between assessment and AI and by focusing on the data available for assessment AI tools. As the American Psychological Association’s (APA) Guidelines for Psychological Assessment and Evaluation (2020) states, “Inasmuch as computer technology, test instrument usage, and new instrument design are constantly evolving, the responsibilities and challenges to the psychologist practitioner using these modalities are likewise substantial.” (p. 22).

It's All Too Seductive
The allure of AI in hiring is undeniable. Algorithms score assessments, generate interview questions, and produce written feedback at a rate far exceeding most HR departments. They promise speed, cost savings, and the promise of objectivity. But the same systems can also discriminate in ways more insidious than human bias. A model trained on data from successful executives might favor middle-aged men with degrees in engineering, thereby excluding women and minorities. Worse, AI can exploit behavioral data—mouse clicks, dwell time, social media activity—of debatable relevance to the job requirements or as a predictor of success, without explicit consent, raising serious privacy concerns. As the authors state, as 'participants' (willing or not) in these solutions, there is a requirement to be better informed how and when these tools are being used. Of course, transparency may not be in a corporations best interest.

Fixing the Machine

Candidates should know when AI is involved, what data are being used, and how decisions are made. Human oversight is essential; AI should assist, not replace, recruiters. Bias audits and retraining of models can help mitigate discrimination, though these require technical expertise and organizational will. The authors advise, consulting psychologists, with their grounding in ethics and statistical methods, are well-positioned to lead this charge. But they must be able to speak the language of algorithms—and to challenge them and the supporters when required.

Prudent Regulation
While the authors may be slightly naive in their advice to cautiously embrace AI since the train is now long gone from the station, they are right that corrective measures must be adopted which are grounded in established psychological theory and ethical standards. Additionally, regulatory frameworks must evolve to keep pace with technological disruption. While no one seems to be saying that AI in hiring is implemented without any human interaction, decisions affecting human lives should not be left to machines alone. Which might seem obvious but that is probably a goal written on a ppt deck somewhere in Palo Alto. Governance comes up a lot when AI is mentioned - possibly because of the lack of it, but as the technology matures, so too must the governance surrounding it. The future of hiring depends not just on smarter algorithms, but on wiser humans.

(Note: I wrote this with help from Co-pilot - like a class assignment. Claude came up with the post title. And CGPT for the image).

APA's PsycArticles contains more than 250,000 full text research articles in psychology but many of these articles have real world business implications relevant to researchers and students in academic disciplines outside psychology and social sciences.  

Anthropic's $1.5 Billion Bargain Bin

On the back of the proposed Anthropic copyright settlement there is almost certainly a long line forming outside the Association of American Publishers (AAP) headquarters to get a piece of the action. In the form of a large copyright penalty discount that is.

Despite the strong objection of Judge William Alsup, who has temporarily rejected the $1.5Billion settlement between Anthropic and a class of authors, this settlement agreement will probably survive. The rejection may appear to delay justice, but it also temporarily protects authors from a deeply flawed agreement that will undermine their long-term interests. While the payout is historic in scale, Judge Alsup rightly flagged concerns about vague documentation, inadequate notice, and the risk of coercing authors into a settlement they may neither fully understand nor meaningfully helped shape. His unease, with the potential for attorneys and organizations to pressure authors reflects a broader fear; that the settlement was more about expediency than equity.  When your professional advocates, in this case, the AAP and Authors Guild (AG) immediately endorse the settlement as a done deal that may not be coercive but it definitely represents pressure.

The case itself is unusually clear-cut. Unlike most AI copyright disputes, which hinge on the unsettled doctrine of fair use, this one involved Anthropic’s admitted use of pirated books from notorious sources like LibGen and PiLiMi. Judge Alsup’s ruling drew a sharp line—fair use may apply to lawfully acquired books, but piracy is indefensible. With an estimated 500,000 infringing works, statutory damages could have reached tens of billions of dollars. Yet the settlement proposed a flat $3,000 per work, a figure that risks – perhaps is already - becoming a de facto benchmark for future cases.  (By the way, fine print suggests the authors won’t actually receive $3,000 per work).

For authors, this would mean accepting a valuation of their work far below what the law allows, and far below what justice demands. Admittedly, in the stark reality of publishing, $3,000 is far more than most of those works are going to generate for their authors, but that is not the point. Copyright law exists to both protect copyright holders (authors) and heavily penalize those who breach copyright. That’s not happening here.

Anthropic’s financial strength further weakens the rationale for settlement. With a valuation projected to reach $170 billion the absurd notion that company is facing a “death knell” from litigation is undermined by the very recent funding rounds topping $10 billion, which proves investors do not consider any financial risks associated with this lawsuit. That should be compelling evidence that this settlement is not even a slap on the wrist.

Approving the settlement will allow Anthropic to sanitize its past misconduct without meaningful accountability, while leaving authors with a diluted remedy.  They should have held out and gone to trial to preserve the only meaningful leverage the authors can exert.

Ultimately, should Judge Alsup reject the settlement only then would this preserve the possibility of a more principled resolution—one that reflects the true value of authors’ work and the gravity of the infringement. Remember, Anthropic (and others) acted knowingly and with impunity to copy these works: With a degree of effort and market knowledge, they sought out these pirate sites precisely because they could get the content for ‘free’.

Rejecting the settlement also avoids setting a precedent that could weaken future copyright claims against AI developers. But while the line outside AAP offices is figurative, I can bet Facebook attorneys are ordering champagne all round this week. 

Authors deserve more than a rushed compromise shaped by corporate interests and vague processes. They deserve a legal framework that protects their rights in the age of machine learning. Until courts provide firmer guidance, the path forward should be cautious, deliberate, and above all, fair. This settlement is not that.

 

****

When the Google copyright case was in litigation, there was a lot of hyperventilation about the universe of 'orphan titles' which, in this highly referenced analysis, I suggested was overblown. We may find that the 500,000 works number cited in this agreement is over egged. 

****

Jimmy tried to force Howard to settle the Sandpiper litigation for all the wrong reasons. It led to Howard's unfortunate demise. 

Clarivate Reports 2Q results - Back on Track?

Clarivate, the information services firm, reported second-quarter results that, while not dazzling, suggest a business regaining its footing. Revenues reached $621 million, and the adjusted EBITDA margin rose to a respectable 41% for the first half—up 500 basis points year-on-year. Organic annual contract value edged up by 1.3%, and recurring organic revenue grew by nearly 1%, buoyed by robust subscription renewals and the rollout of new offerings.

The headline net loss of $72 million was largely the result of non-cash impairment charges, and on an adjusted basis, net income stood at $123 million. The firm announced they will seek to return up to $100 million to shareholders via buybacks and reiterated its full-year free cash flow guidance of $340 million.

Strategically, Clarivate is leaning into its strengths. Its Academic & Government and Life Sciences & Healthcare segments showed momentum, aided by AI-driven innovation and fresh partnerships. The Intellectual Property division, long a laggard, returned to growth—thanks in part to a surge in AI-related patent filings, especially from China.

Challenges remain. Divestitures and a sluggish commercialization market continue to weigh but management remains positive, touting its Value Creation Plan and a revamped sales model. The company reaffirmed its 2025 outlook, projecting revenues between $2.28 billion and $2.40 billion, and adjusted EBITDA between $940 million and $1 billion.

In short, Clarivate is not yet sprinting—but it is walking with purpose.

*********

The following are some comments from the investor call transcript. These comments are about the previously announced switch from a book purchase to book subscription model which generated a lot of push back from librarians (and other business initiatives):

“If you turn to Slide 8, I'll provide an update on the VCP starting with the A&G segment. Our proactive business model optimization, coupled with decades of experience in delivering data and analytics solution to our clients has strategically positioned us to anticipate and adapt to current market dynamics. We are on track to discontinue transactional sales of digital collections and books over the next year. This shift away from transactional sale is increasing recurring revenue growth by transitioning some of the business to the new progress e-Books product and other content solution subscription. We are pleased with the early adoption with over 70 wins to date and hundreds of customers currently evaluating this new model.”

“Following this A&G strategy, A&G subscription revenue now constitutes 93% of the total segment revenue, excluding disposals, up from 79% in the prior year period. In the first half of 2025, we have achieved a 96% renewal rate in A&G. This is impressive results considering the macro backdrop characterized by a reduction in the U.S. federal agency contracts, increased constraints on higher education research funding and potential additional university budget cuts. It is also noteworthy that as at the end of July, 75% of global A&G subscription for the full year has successfully renewed. This is in line with last year -- last year's renewal pace. We continue to successfully invest in innovation across the A&G product portfolio with a focus on AI. We are very pleased by our success so far in product launches and customer adoption. More than 4,800 institutions have already adopted our AI tools to strengthen research support, increase operational efficiency and enhance student engagement.”

“We are also accelerating progress with next-generation Agentic AI solution. AI agents can independently play and execute multistep processes by interacting with user with users, data sources and tools. The expansion of our Agentic AI platform marks a significant milestone as we implement responsible Agentic AI to accelerate research and learning workflow. Our initial launch of the literature review agent in Web of Science exemplify this pioneering approach. The agent converse with researchers to understand their research goals, then customize a specific literature review scope and define the proper output. This personal interactive experience keeps the researcher in the center, which closely mimic working with human assistant.”

“With regards to the content, as I mentioned in my discussion, we were forward looking, taking away the discretionary onetime expenditure, and this served us very, very well these days, and we see the uptake of customers who were initially complaining about taking away the onetime purchases are now buying more and more of our subscription businesses, the PQ ebook, the PQ Digital Collection, which actually serves them very, very well in this in this kind of economic climate. So we are pretty confident that going ahead, we will continue to see a good and decent renewal rates and uptake of the different A&G offering.”

Wiley Full Year Results Show Margin Improvement

For Wiley's Fiscal Year 2025 (ending April 30, 2025), the fourth quarter reported revenue was $443 million, essentially flat year-over-year when adjusted for divestitures, while full-year adjusted revenue was also essentially even. The company saw growth in its Research and Learning segments, particularly in academic courseware and AI licensing, but experienced softness in professional sales due to a weak retail market. Full-year adjusted EBITDA grew to $369 million (at constant currency), driven by cost savings and revenue performance in Research.

Wiley continues to experience robust growth across its Research and Learning segments with recurring revenue performing well, supported by strong research submissions and output. The India, Brazil, and China markets are showing significant revenue improvement and the business unit is seeing double-digit submission increases from the UK, France, Italy, and Canada. The Japan and the U.S. markets are contributing with high single-digit growth.

Open access publishing remains a key strategic initiative, with double-digit growth over the year. AS noted in previous announcements, Wiley has made notable strides in AI licensing, securing its third major AI model training customer in Q4 and generating $40 million in AI-related revenue for FY25. The company is expanding its corporate sector footprint through vertical-specific licensing agreements with technology, pharmaceutical, chemical companies, and a space agency. Strategic partnerships with Amazon Web Services and Perplexity are enhancing Wiley’s capabilities in research API development and AI integration.

In Higher Education, Wiley is seeing strong performance in digital and AI-driven offerings. AI licensing revenue reached $29 million, up 27% year-over-year, supported by both Academic and Professional backlists. Inclusive Access programs show strong double-digit growth and academic digital content and courseware are growing steadily.

Wiley is also innovating in assessments and team development, launching new products that combine personality models with training solutions. The company believes their thought leadership in AI development, research integrity, and accessibility positions them well for growth and an expanded presence in both the academic and corporate sectors.

More information from their press release:

FISCAL 2025 HIGHLIGHTS

• GAAP performance vs. prior year: Operating Income of $221 million vs. $52 million and Diluted Earnings Per Share (EPS) of $1.53 vs. ($3.65)
• Exceeded Adjusted EPS guidance, delivered at top end of range for Adjusted EBITDA margin, and achieved Free Cash Flow outlook
• Delivered Revenue and Adjusted EBITDA margin growth in both Research and Learning segments
• Achieved Adjusted Operating Margin expansion of 300 basis points
• Executed AI content licensing project this quarter with a third large tech company; $40 million in total AI licensing revenue realized in Fiscal 2025 compared to $23 million in Fiscal 2024
• Drove a 34% increase in share repurchases and raised dividend for 31 ^st consecutive year

Ava Gardner Beat Artie Shaw at Chess: Ava The Secret Conversations

Elizabeth McGovern, Photo by Jeff Lorch

Ava Gardiner didn't like a woman to swear. She fucking hated it. In Ava: The Secret Conversations Elizabeth McGovern embodies the life of Ava Gardner as her life story is coaxed out of her by writer Peter Evans (played by Aaron Costa Ganis) over the course of several weeks in 1988. Evans is badgered into getting the salacious out of Ava by an audio only Ed Victor who notes frequently that no less than Dick Snyder himself is paying attention to the writers' efforts. 

The play opens with a 2am call to Evans by Gardner where she seems to be contemplating suicide - and although this doesn't come up again in the play, the reason she has entertained the idea of selling her story at all is that she needs the money for her mortgage and is still suffering from the effects of a stroke. She's not in a good place, and she will question some of her life choices during these conversations. McGovern plays the role with her left arm mostly hanging limp except when the scenes fall back to reenact past events. As these events unfold the stage scene is boosted by the use of video images of past husbands Mickey Rooney, Artie Shaw and Frank Sinatra. Ganis will later break out into song.

It is Frank, that Synder desperately wants in this book and he presses, via the voice of Ed Victor, Evans to get the anatomical bits out of Ava. At the time, Sinatra had recently been the victim of the Kitty Kelly bio and Snyder knows how much anything about Sinatra would drive sales of this book. Even without Frank - who does feature, there is much to consider here about Ava Gardner: McGovern and Ganis are galvanizing in their respective roles. The set which represents Gardner's apartment in London is both sitting room and bedroom - the bed being obviously symbolic, is very well constructed and the multimedia images add significant context to the story.

As it turned out, Gardner threw Evans out before the manuscript was finished and it was only after she died in 1990 that Evans revisited the material and with the approval of the Gardner family the work was published. From this book, the play was conceived by McGovern and premiered at the Geffen Playhouse in 2023.

And yes, on the first occasion she did beat Artie Shaw at chess even though he was an expert. Go see it.

***** 

For those who don't know, Dick Synder was the long time CEO of Simon & Schuster and by accounts liked nothing better than his executive team to be in open combat with each other. Ed Victor, was a 'super-agent' sometimes known for influencing the telling of the story. 

 

Building an Intelligent Supply Chain in Publishing

The Book Industry Study Group is embarking on a Book Publishing Next initiative to rethink the publishing supply chain/ecosystem. See more at https://lnkd.in/eqemKjMW

It's a long time coming: Here is my thoughts on this from 2002!
https://lnkd.in/et3F6wPS

 

 

Building an Intelligent Supply Chain Frankfurt Supply Chain Interests Group 2002 from Michael Cairns

The Satire Bites...and Leaves a Mark

Don't we just love watching The Daily Show and Late Night and the skewering which takes place all in the name of fun? Sometimes this satirical "fake news" can cascade into the scatological - particularly The Daily Show commentaries, and new research published in The Journal of Experimental Psychology (and available via PsycNet subscription) demonstrates that satire can pack a punch and will frequently result in the dehumanization of the target.

Long considered the harmless cousin of criticism, satire may in fact be the more dangerous sibling. A recent study entitled Softening the blow or sharpening the blade: Examining the reputational effects of satire. by Hooria Jazaieri (Santa Clara University) and Derek D. Rucker (Northwestern University) suggests that satire, despite its comedic veneer, inflicts greater reputational damage than direct critique. The authors set out to test two hypotheses: Does satire soften the blow or does it negatively sharpen the impact. Their findings favor the latter.

The study examined how satirical content—memes, videos, and commentary—affects public perception. Participants exposed to satirical portrayals of individuals, both real and fictional, consistently rated those individuals less favorably than those exposed to straightforward criticism. Satire, it turns out, does not merely mock—it dehumanizes.

Read a profile of Stewart in a recent Poynter article - When Satire became trusted news.

A linguistic analysis of more than 100,000 YouTube comments revealed that satire prompted viewers to use language that stripped its targets of human qualities. This dehumanization, measured using the Mind Attribution Scale, was found to mediate the reputational harm. In short, satire makes people seem less capable of thought, emotion and agency—and thus less worthy of respect. (Note: The Mind Attribution Scale (MAS) is a questionnaire designed to measure the extent to which individuals attribute mental states, specifically agency and experience to others). 

To determine whether there exists a counter to this impact, the researchers tested a psychological intervention: asking participants to imagine a pleasant interaction with the satirical target. This brief exercise, lasting no more than two minutes, significantly reduced the reputational damage. The imagined contact restored perceptions of the target’s humanity, blunting satire’s sting.

So, consider this 'mission accomplished?' In an age where satire proliferates across social media and entertainment platforms, its reputational consequences are often overlooked - although, some could also argue undermining reputationally and otherwise is the objective. The researchers argue that far from being a benign form of commentary, satire is a potent weapon—one that shapes public opinion more effectively than direct criticism. The ethical ramifications are considerable, particularly when satire targets individuals unable to respond or defend themselves.

APA's PsycArticles contains more than 250,000 full text research articles in psychology but many of these articles have real world business implications relevant to researchers and students in academic disciplines outside psychology and social sciences.