The Future Of Open Access In Science

$3,000 open-access fee isn't just steep - it's creating a new form of academic inequality in 2025. In the AI era, if your research isn’t open access, it may not even exist. Not to AI. Not to the next generation of researchers. Not to anyone relying on tools like Elicit, Scite, or Scispace to surface the literature. That’s the reality we’re heading into. And we’re not fully ready for it. Here’s what’s happening 👇 AI tools mostly read what’s publicly available online. They prioritize open access by default—not by principle, but because that’s what they can see. Closed-access papers? Many of them are invisible to AI. Sure, newer agentic AI (like Deep Research) now bypass paywalls or access shady archives. But those tools aren’t mainstream (yet). Most don’t have access to them. And that creates a quiet split in visibility: → Those who can afford open access = more citations, more AI reads, more impact. → Those who can’t = invisible. That’s the part that worries me. We’ve seen this before. Visibility tied to privilege. Access tied to resources. And now, AI might amplify that divide. 7 strategies to tip the balance in smart, ethical ways: 1️⃣ Preprint Servers (Gold Standard for AI Access) 2️⃣ Conference Archives That Host Full Papers 3️⃣ Institutional Repositories 4️⃣ Government-Funded Collaborations 5️⃣ Attach Full PDFs to Research Profiles 6️⃣ Publish in Smarter OA Journals 7️⃣ Use Structured Metadata Will AI finally push the big journals to go fully open access? Maybe. But until then, we’re in a transition phase. And in this phase, discoverability isn’t just about what you write. It’s about where and how you publish it. So make it count. Because in the AI world, being invisible isn’t a reflection of your work’s value—it’s just a technicality. One you can control. What are some other strategies you use for increasing visibility?- --- P.S. Join my inner circle of 5000+ researchers for exclusive, actionable advice you won’t find anywhere else HERE: https://lnkd.in/e39x8W_P BONUS: When you subscribe, you instantly unlock my Research Idea GPT and Manuscript Outline Blueprint. Please reshare 🔄 if you got some value out of this...

ArXiv and open access articles are cited much more by AI Like most professors, when I'm not complaining about AI, or writing articles about the problems caused by AI, I'm using AI as part of my work. Recently, I had an interesting experience in attempting to survey a new (to me) research area related to some WiFi implementation issues. I did a series of searchers using the new GPT 5. Of course, I'm also doing a review of the literature in other ways as well. I have found AI based searching to be one of many possible tools to use. During the searching, I asked the AI to give me references that I could cite with appropriate links. All the references I got were to technical preprints that were on the arXiv preprint server, another preprint source like ResearchGate or had appeared as open access often in pay-to-publish MDPI journals. Papers that appeared behind the paywall in IEEE Xplore, where I usually publish, did not appear. I tried to get GPT 5 to give me some links to papers from IEEE Xplore, but it indicated it was blocked from searching there. It did give me some publications from arXiv that may also be found on IEEE Xplore. What does this mean? Publications that appear in what are considered the most prestigious IEEE journals (at least in my research area) are locked behind a paywall. As people increasingly use AI to help in literature reviews, these ``prestigious'' papers will get overlooked in favor of non-peer reviewed papers or open access papers. For researchers who still want visibility, the easiest option is to post preprints on arXiv or a similar online sources. Another possibility is to choose an alternative quality option (if available) that will result in something that gets crawled by the AI engines. A final option is to is to pay the high fees to make a paper open access so that it doesn't live behind the paywall (though I'm not totally sure this will work in all cases). Long term, this is yet another crack in the conventional publishing model. We authors are already asked to pay overlength page-charge fees, despite that submitting a paper less than the maximum length is less likely to be accepted. We are expected to review for free. We have to deal with increasingly noisy or nonsensical reviews and decisions by editors that don't read the paper. Funding is tight (in the US) making it hard to fund additional open access fees on top of the overlength page charge fees. It might be a good time to strategize again about how and where we publish. 

Interesting and exciting developments in open science! A recent Nature article highlights an innovative initiative from CERN to financially incentivize publishers that adopt open science practices. Under the new Open Science Mechanism, journals can earn higher payments by embracing practices like transparent peer review, linking research to datasets, and making content more accessible. This is a bold step forward, recognizing the values of openness and transparency in advancing research. As Kamran Naim, PhD of CERN aptly said, “Openness is the only really effective way of practicing science” in complex, global disciplines like particle physics. CERN’s model is a game-changer: it’s tying financial rewards to measurable progress in open science, and working with publishers who, are the gate keepers of quality, and can make interventions with authors at the right time in the publication workflows. This actually resonates deeply with what we’re doing at DataSeerAI and our Snapshot product — helping publishers, funders and institutions take meaningful steps towards publishing transparent, open and reproducible research. As the open science movement grows, initiatives like CERN’s—and tools like DataSeer—are paving the way for a more equitable and collaborative research ecosystem. #OpenScience #OpenData #ScholarlyPublishing #CERN https://lnkd.in/ecJAFE_v