THE TOLL FOR TRUTH · PART 7 OF 7
Over the last six posts, I’ve argued that the paywall is a rational choice for each publisher that adds up to an irrational outcome for everyone. It sorts readers by ability to pay rather than need, leaves news deserts unfilled, charges the public twice for research it funded, and cedes the open web to whoever can afford to be free.
If you’ve been reading along, you might expect this post to say “tear down the paywalls.” It won’t. The paywall exists because the ad-supported model broke, and pretending otherwise doesn’t help anyone. No single fix will solve this. But the evidence does point toward five complementary moves, each aimed at a different point where the system currently sorts people out.
Each recommendation targets a different problem and a different set of actors. (Figure 14 from the white paper.)
1. Extend and measure public access policy
The 2022 OSTP memo and NIH’s 2025 policy were real wins. But they cover only federally funded research, they require only the author’s accepted manuscript rather than the final version, and they don’t touch the backlist or research funded by states, philanthropies, or international sources.
Future policy should extend immediate public access to research supported by state and philanthropic funders, and encourage depositing final published versions wherever publisher agreements allow.
Just as important, policy needs evidence. There’s currently no baseline data on what unaffiliated and independent researchers spend to access the literature, or how often paywalls delay or end their work. Without that, policymakers can’t tell whether open access efforts are actually reaching the taxpayers they’re meant to serve.
2. Resource libraries to serve people outside the academy
Public libraries are the natural front door for researchers without an affiliation, but many can’t afford the specialized databases those researchers need. It doesn’t have to stay that way. Efird-Green and colleagues describe how one institution rebuilt a research e-library for community users, including checking whether each resource was actually usable without an institutional login.
That approach could scale through statewide consortium licensing, community borrower and alumni access programs at academic libraries, and dedicated funding for public library database subscriptions. Libraries are also well placed to collect the usage and demand data that would finally make unaffiliated researchers visible to publishers and policymakers.
3. Turn the crisis exception into a standing civic tier
Publishers already drop their paywalls for elections, disasters, and public health emergencies. Why not make that the default? Reporting on elections, public health, and public safety could be free by standing policy rather than by emergency declaration. Publishers keep subscription revenue for the bulk of their output, while citizens get unrestricted access to the information they most need to take part in public life.
Pricing can help too. The Reuters Institute found that a fifth of non-paying Americans are interested in a reasonably priced bundle covering multiple news providers. Bundles, income-based pricing, and library-mediated news access are all ways to lower the toll without giving up reader revenue.
4. Hold AI systems accountable for access and accuracy
AI developers now stand between a lot of readers and the verified record. They should share both the costs and the accountability. That means three things:
- License and compensate the publishers whose work their systems summarize.
- Attribute accurately, with working links to original sources.
- Hold free tiers to the same accuracy standard as paid ones. As the EBU and BBC evaluators put it, access to accurate news shouldn’t depend on whether you pay for an AI assistant.
Regular public reporting of error rates, broken down by product tier and language, would let users, publishers, and regulators see whether these systems are actually getting better.
5. Demonetize the Mass Distribution Loop
Fixing the economics of journalism is only half the job. The economics of misinformation need attention too. Ahmad and colleagues found that most misinformation sites run on advertising, and that most company decision-makers don’t know their ads appear there.
Transparency attacks the loop right where it turns reach into revenue. Ad platforms could disclose where programmatic ads are placed, and advertisers could get easy tools to exclude sites that repeatedly publish false information. Shrinking the subsidy for falsehood narrows the gap between the free and verified tiers from the other direction.
What this series can and can’t claim
I want to be upfront about the limits here. The white paper synthesizes secondary sources; it isn’t an original empirical study. Some of the figures on AI crawling and content reconstruction come from industry sources, including vendors with commercial interests in the problems they describe, and should be read as estimates. Much of the data is U.S.-centric, and the AI landscape is moving faster than the research can follow. Several claims about what unaffiliated researchers pay rest on inference from affiliated researchers, because those costs have never been measured directly. That gap is the one I most want to see closed.
The bottom line
Paywalls emerged to preserve the institutions that produce verified knowledge, and in many cases they have. But a society that preserves its sources of truth by pricing most of its members out of them hasn’t solved the problem of an informed public.
It has privatized it.
The toll for truth is being paid by the people least able to afford it, in money, in time, and in the quality of the information they’re left with. Deciding who should bear that cost is ultimately a public question, not a market one.
Thanks for reading the series. If you want the full argument with every citation, the white paper is linked below. I’ll dig further into the AI side of this story in future posts.
Read the full paper: This series is adapted from Konderla, J. (2026). The Toll for Truth: Paywalls, Information Deserts, and the Asymmetric Web (Version 1.0) [White paper]. Zenodo. https://doi.org/10.5281/zenodo.23131192 The paper includes the complete argument and full bibliography.
Sources
Ahmad, W., Sen, A., Eesley, C., & Brynjolfsson, E. (2024). Companies inadvertently fund online misinformation despite consumer backlash. Nature, 630, 123–131. https://doi.org/10.1038/s41586-024-07404-1
Efird-Green, L., et al. (2024). Designing and implementing a community-engaged research e-library. College & Research Libraries, 85(7), 994–1005. https://doi.org/10.5860/crl.85.7.994
European Broadcasting Union & BBC. (2025). News integrity in AI assistants. https://www.ebu.ch/Report/MIS-BBC/NI_AI_2025.pdf
Galter Health Sciences Library. (n.d.). NIH public access. https://libguides.galter.northwestern.edu/NIH-public-access
Reuters Institute for the Study of Journalism. (2025). Digital news report 2025. https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2025/dnr-executive-summary
White House Office of Science and Technology Policy. (2022). Ensuring free, immediate, and equitable access to federally funded research. https://bidenwhitehouse.archives.gov/wp-content/uploads/2022/08/08-2022-OSTP-Public-Access-Memo.pdf