Adapt, adopt, assimilate—that is the way digital solutions vendors are going about making technologies, new and proven, work for them and their publishing clients. This method also applies to their stance on three hot industry topics: accessibility, artificial intelligence, and automation.
Improving efficiencies, reducing production costs, and shortening turnaround times for deliverables equal higher profit margins and faster time to market. These advantages benefit both vendors and clients, allowing them to broaden their services and enter new product segments. After all, who is not paying close attention to the massive scale of the global digital publishing market and seriously contemplating having a bigger cut of the potential business?
Take the latest SkyQuest report, for instance: it predicts the market will expand from $186.77 billion in 2023 to $367.19 billion by 2031, growing at a CAGR of 9.67% over the period.
According to the same report, video is set to be the fastest-growing content type, followed by text and audio. Another report from Technavio states that the digitization of paperback books will boost the digital publishing market, which will increase by $138.76 billion between 2024 and 2028.
Of course, the numbers differ from report to report, but the growth and potential is undeniable regardless of where one looks. And this potential is driving the never-ending quest for more efficient and seamless workflows, innovative technologies and tools, and robust processes, which comes with a lot of experimentation and plenty of risk. But then again, no successful business comes without considerable risk-taking. For this report, PW talks with two digital solutions vendors—one based in Chennai and another in the U.S. with production facilities in India—to get a quick update on their companies while exploring hot issues affecting the industry in general.
Adding and adopting AI
Take a gander at these recent headlines: AI-written articles under fake bylines. AI-based plagiarism is on the rise. AI paper mill scandal exposed. It is no wonder alarm bells are going off about the use of AI in the publishing community.
But help is on the way: governments and legislatures are stepping up to exert some control over AI implementation and usage in various sectors. The groundbreaking EU Artificial Intelligence Act, for instance, will be enforced by the end of July 2024 to “foster trustworthy AI in Europe and beyond, by ensuring that AI systems respect fundamental rights, safety, and ethical principles and by addressing risks of very powerful and impactful AI models.”
Currently, in the content and publishing space, the use of AI-driven tools to speed up the production process is a reality. Mundane and tedious copyediting, glossary-making, proofreading, and indexing are becoming faster and more efficient with the help of AI-driven automation tools. Harnessing AI and balancing it with human subject matter expertise is the way to go, and this combination will benefit both publishers and digital vendors.
The Westchester Publishing Services team, for one, is definitely excited about AI. Nevertheless, “the risk of content piracy and scraping,” says chief revenue officer Tyler Carey, “is making us very selective with the models we are exploring. Through dialogue with clients on AI, I would say that one-third is cautious, one-third ambivalent, and the rest are hoping that we can bring AI into their workflows to help speed up production and reduce costs. That seems like a sensible—and practical—mix to me, rather than, say, when AR or blockchain became buzzwords that were too costly, impractical, or confusing to adapt for many publishers. Our own explorations are focusing on models that are closed-off walled gardens to avoid the aforementioned risks.”
Keith Riegert from Ulysses Press and Perfect Bound, Carey says, “presented a wonderful overview of different platforms at the U.S. Book Show in May. On the PerfectBound.io site, he rounded up a number of use cases that are worth checking out, and I am excited to see the other platforms popping up that have some real potential for publishers for organizing their data and content.” In particular, Carey finds Dropbox Dash and Notion AI to be exciting platforms for publishers to explore. “The Book Industry Study Group (BISG) has put together an AI working group where publishers are regularly sharing ideas, and I encourage publishers to check it out to learn about how their peers are using these tools.”
For Carey, AI at Westchester can mean lots of things. “Dave Cramer from Hachette and other thought leaders have spoken about true AI vs buzzword AI, the latter referring to AI that is more mundane—but valuable—like grammar or spell check,” Carey says. “At Westchester, we have been using more mundane AI such as a natural language processing code of our own design to create keywords for client marketing metadata and non-GPT tools to standardize citations, for instance. OpenAI models for client content have risks that we are avoiding by using alternate models that provide the right functions and security levels.”
Over at Lapiz Digital, company president V. Bharathram finds that publishers are becoming more interested in using AI tools and services to improve content production and delivery. “Most of our projects are with educational publishers, and generative AI has been a great help and has provided a range of assistive capabilities for SMEs and content developers,” Bharathram says. “We have worked with AI-enabled content-related projects for K–12 clients, and the majority of these projects involved creating assessment items for various learning levels in different subjects.”
However, Bharathram and his team have noticed that when they used AI to simultaneously generate numerous items from a test content area, duplicate content was often produced. “These items were also uneven in their coverage of topics within that content area,” Bharathram says. “Then there are concerns about bias, discrimination, and ownership in AI applications. But we are confident that everything will eventually evolve for the better. Until then, SMEs and experienced content developers will continue to review AI-generated content before it is published—and Lapiz Digital is here to help them along the way.”
Analyzing and enhancing accessibility
For the estimated 1.3 billion people with significant disabilities, assistive technologies for digital content are enhancing access at a fast clip and moving way beyond color-and-contrast and text-to-speech solutions. Consider the latest Apple announcement on its upcoming accessibility features, which are mostly powered by AI: Eye Tracking, which enables users with physical disabilities to navigate iPad or iPhone with their eyes; MusicHaptics, which enables those with hearing impairments to experience music; and Vocal Shortcuts, which allows users to perform tasks by making a custom sound. Then there is Hover Text, which shows larger text above a text field when a user is typing to support those with vision issues, and Listen for Atypical Speech, which helps people whose speech patterns diverge from the norm.
Regulations in Europe and the U.S.—the European Accessibility Act (EAA) and the Americans with Disabilities Act (ADA), respectively—are increasing awareness of accessibility and, by default, of diversity, equality, and inclusivity as well. But the progress on ensuring digital accessibility has been bumpy and slow. Issues such as videos lacking closed captions, non-descriptive error alerts, websites that are unnavigable via keyboard, and unapproachable screen readers are persistent and prevalent. And it is not just about making videos or websites accessible for people with auditory, cognitive, motor, and visual disabilities. Emails, PDFs, and PowerPoint presentations also need to be made accessible.
For digital vendors such as Lapiz Digital, EAA 2025 is a transformative directive that will redefine how businesses operate, innovate, and engage with consumers. “Though we are not seeing a major increase in accessibility projects at the moment, we are sure this will change in the near future,” says Bharathram, whose company is certified for accessibility production services by Benetech. “Thus far, the accessibility projects coming through our doors have been limited to K–12 products.”
Publishers, Bharathram says, have different criteria for accessibility based on a number of variables, including the type of content they generate, the audience they cater to, and the jurisdiction in which they operate. “Our team provides accessibility support for both graphics and alt text,” he says. “In one recent project, we had to create digital copies of a client’s products using EPUB3 standard’s accessibility capabilities along with the product descriptions.”
Accessible EPUBs can, and should, be affordable and require nowhere near the spending on copyediting or designing that a print book does, says Carey, of Westchester. “As a Benetech GCA-certified vendor, we have been trying to educate the market not just on what is coming with the European Accessibility Act (EAA) but on how to incorporate these needs into the workflow right now.” For more on accessibility, see Carey’s online article “Navigating Approaching Accessibility Requirements,” which is adapted from his blog post about best practices for BISG.
But for publishers, bringing their backlists into compliance with the EAA is not easy. “Publishers have realized that they have a shrinking timeline to address this, and the sheer volume of published titles they have to remediate is not insignificant,” says Carey, whose clients are trying to figure out how to budget for this requirement in the near future. “The remediation process is not a challenge for Westchester—we are here and able to help— and our offerings are affordable. However, working on backlists involves a much tougher conversation about incurring a production spend after a book has already been in the market, rather than a more typical conversation around the budget for editing or typesetting a book that has yet to be released and have all of its costs recognized.”
Thinking back and moving ahead
Remember Plan S? This open-access (OA) publishing initiative for scientific publications caused an uproar among authors, publishers, and researchers when it launched in Europe in September 2018. Soon after, the pandemic hit and the ensuing race to secure medication and vaccines highlighted the importance of having OA shared research and results. And in August 2022, the White House Office of Science and Technology Policy (OSTP) issued a new OA policy similar to Plan S for federally funded research. The push for OA publications went into overdrive.
Presently, funders of Plan S—the international consortium known as cOAlition S—have expanded to include organizations such as the Bill & Melinda Gates Foundation and the World Health Organization. The bad news is that cOAlition S funders are ending their financial support of OA publishing under transformative arrangements—that is, moving from subscription- or hybrid-based publication to full OA—after December 2024.
According to a recent paper written by Robert-Jan Smits, president of Eindhoven University of Technology and former director-general of the European Commission, who developed Plan S, about 61% of the four million scientific papers published annually are still behind subscription paywalls. In critical medical research areas such as cancer, cardiovascular disease, and respiratory disease, only 31%, 20%, and 16%, respectively, are open access. He also says that in the field of climate change, which poses a major challenge to our very existence, OA publications represent only about 40% of the literature.
Whether Plan S will continue to gain momentum is anyone’s guess. Recent months have seen the funders pushing for a more open, equitable, and faster publishing system for scientific research in which authors decide where and when to publish their work. This scholar-led proposal does not seem popular among the publishing community. Upending the subscription business model is iffy at best, especially with publishers fighting to maintain their bottom line while making new research available and investing in technology solutions and new products. Major institutions are also encumbered by long-term journal subscriptions that are not easy to scrap. For digital solutions vendors, the mission remains unchanged: lower the production costs and processing time of OA articles and journals to help make the Plan S proposition viable for publishers.
Then there is the General Data Protection Regulation (GDPR), the EU’s flagship data protection law. Celebrating its sixth anniversary in May, it has set a global benchmark for regulators in terms of reshaping data management and prioritizing privacy rights.
But how does the GDPR jive with the EU Artificial Intelligence Act, in which human oversight is required for certain AI systems such as large language models (LLMs) that may contain sensitive private information or datasets collected by third parties? Obtaining informed consent and managing data subject rights under this model would become so complex. AI-driven large datasets and algorithms, which are prone to bias, pose additional challenges to key GDPR principles of transparency and data minimization. So opinions are split across two opposing camps: either GDPR is too restrictive to allow AI to flourish or it is too ineffective to properly regulate AI.
Enter the German federal and state data protection authorities, which recently published guidelines on the implementation and use of AI in compliance with GDPR. These guidelines primarily address deployers of AI applications and identify some risks of AI use, including unlawful processing of personal data and discrimination through biased data. They are in line with Article 50 of the EU AI Act, in which deployers and providers must state clearly whether an individual is interacting with chatbots or whether a specific piece of content is AI-generated, for instance.
This leads us to the topic of metadata, which has been all but drowned out by the talk around accessibility, AI, and automation. Nevertheless, publishers and digital vendors are still paying attention to the who, what, where, when, why, and how of an article or a book or journal. Getting the metadata right means getting the content optimized for sale, which the publishers do get, pun intended.
At Westchester, for instance, metadata remains key. “Our most direct interactions with metadata are through enrichment by adding keywords, generating BISAC codes, and inserting accessibility metadata,” Carey says. “Discoverability is of utmost importance to publishers, and metadata is the crux of that.”
Then there is blockchain, a hot topic at every book event prior to the pandemic. “Blockchain is fascinating—I would put in the same category as AR, VR, and MR—and it has real potential, but it is just not something I see most publishers really pursuing,”
Carey says, adding that Book.io is a possible exception. “I am curious to see where it goes with blockchain applied to EPUBs. But I think conceptually, blockchain is harder for many publishers to wrap their head around from an implementation standpoint versus, say, implementing AI as part of their processes, since the barrier to entry to adapting AI is so much lower.”
Meeting new demands and trends
While digital solutions vendors may dream up infinite tweaks to their technologies and tools, human resources, unfortunately, are finite. Hiring and retaining skilled software experts to keep abreast of technological advances is a big challenge, says S. Nagarajan, operations director at Lapiz Digital. “We have sketched out plans for NLP- and AI-integrated automations for the next two years, and finding suitable software experts with relevant publishing industry experience is a major task. At the same time, ongoing simplification of production processes via technological advances is often accompanied by a price crunch from clients. But we will continue to improve our skills and offerings, as well as our profit margin.”
Continuous training is par for the course. “Our team is trained in various in-demand, content-related services as well as in areas where we anticipate future advances and demands,” Nagarajan says. The latter covers multimedia content creation, accessibility-related offerings, advanced animations with JSON, and assisting publishers with educational strategies including content gap analysis and learning pathways design.
Bharathram, on the other hand, anticipates rapid development and more changes in edtech and LMS in the coming months, “as long as educators are looking for creative methods for improving teaching and learning opportunities and as technology keeps advancing,” he says. The Lapiz team continues to receive projects in areas such as assessment development, supplement authoring, pedagogical reviewing of AI-generated content, and content validation against assessments. “Our proficiency in specific tools—Canvas, GeoGebra, MATLAB, Octave, and SoftChalk, for instance—ensures that we are the top-tier content partner for many edtech companies,” Bharathram says.
In today’s ever-changing publishing environment, publishers that follow the trends, value diversity and inclusivity, and are willing to experiment with new digital content formats stand a better chance of success and survival, Bharathram says. “Immersion-based and interactive content experiences are drawing in more readers. So publishers are experimenting with enhanced e-books, interactive storytelling concepts, and VR experiences in an effort to engage readers in novel and creative ways. Plain-text e-books without additional multimedia elements are no longer as popular as they were, say, a few years ago.”
Experimentation with new formats, such as blockchain, AR, VR, and MR, however, are often accompanied by high production and technology costs. Alas, cost pressure is a harsh reality for both publishers and vendors. Automation and improved work efficiencies are driving down costs, but there is only so much to shave and save before you hit the floor. Now is the time, yet again, to revise the digital solutions playbook, add new rules—especially those concerning accessibility, AI, and automation—and get on with the game. Remember, fortune favors the bold!
