APIs are essential infrastructure for research, and the politics of who gets API access for research purposes is one of the quieter but more consequential fights in the API economy. Academics, scientists, journalists, and independent researchers depend on APIs to collect data, study platforms, audit algorithms, and understand the systems that increasingly shape society. When a researcher studies political discourse on social media, investigates how a recommendation algorithm behaves, or analyzes patterns in public data, they often do it through APIs. This makes API access a precondition for a huge swath of modern research — and it makes the platforms that control those APIs into gatekeepers of what can and can’t be studied. The politics of research access to APIs is, at bottom, about whether the systems that govern our lives can be independently scrutinized, which is a question with profound implications for accountability, science, and democracy.
The positive role of APIs in research is real and worth celebrating, because APIs genuinely accelerate science. I wrote in 2016 about making scientific research more real-time and collaborative using APIs — using the Zika virus research as an example of how APIs let researchers share data and collaborate in ways that speed discovery. APIs are how scientific data flows between institutions, how research datasets become accessible, how collaborative science happens across distance. I’ve written about the question of whether API platforms can ensure access to federally funded research, about keeping important scientific research projects alive through API access, and about university API efforts. When APIs work for research, they’re a genuine accelerant — democratizing access to data, enabling collaboration, and letting researchers build on each other’s work. The vision of open, accessible, API-driven science is one of the more hopeful applications of everything I’ve evangelized.
The fragility of research access is where the politics turns dark, because research depends on access that platforms can revoke at will. Researchers who build their work on a platform’s API are in the same precarious position as any other API consumer, but with higher stakes: a doctoral student whose dissertation depends on Twitter data, a research team studying disinformation through Facebook’s API, an investigator auditing an algorithm — all of them are exposed to the platform’s power to cut off access. I wrote in 2017 about the need for service-level agreements for researchers who depend on APIs, and about having a program for researchers baked into your API operations — because researchers need stability and assurance that the access they depend on won’t vanish mid-study. The reality, though, is that research access is frequently treated as expendable, and when platforms lock down their APIs, researchers are among the first cut off. The many ways in which APIs are taken away, which I catalogued in 2019, hit researchers especially hard, because their work can’t be quickly rebuilt on a different platform.
The platform-lockdown crisis devastated research access, and it’s one of the clearest cases of platform power harming the public interest. When Twitter, Facebook, and other platforms restricted their APIs — Twitter most dramatically in 2023, pricing API access into the stratosphere — they cut off a generation of researchers who had been using that access to study the platforms themselves. This is the cruelest irony: the researchers most in need of API access are often the ones studying the platforms’ own effects on society — disinformation, political manipulation, algorithmic harm — and they’re exactly the ones the platforms have an interest in cutting off. I wrote about Twitter as the most important API in part because of its role in research, journalism, and understanding democracy, and its lockdown was a genuine loss to the public’s ability to understand what was happening on the platform. The Cambridge Analytica fallout made this worse, because platforms used “protecting user data” as cover to restrict the legitimate research access that would let independent scholars hold them accountable.
The algorithmic-auditing dimension is where research access becomes most politically charged, and where I’ve pushed hardest. I argued in 2016 that if an algorithm impacts our lives, it should be opened up with an API for auditing — so that researchers, journalists, and regulators could independently scrutinize the algorithms that increasingly govern consequential decisions. This is research access as democratic accountability: the algorithms that shape what we see, how we’re treated, and whether we’re policed should be auditable, and APIs are a mechanism for that auditing. But the platforms whose algorithms most need scrutiny are precisely the ones least willing to provide research access to them, because algorithmic transparency would constrain the power that algorithmic opacity protects. The fight over whether researchers will have the API access needed to audit consequential algorithms is one of the most important political fights in the entire field, because it determines whether the most powerful and least visible systems in modern life can be independently studied and held accountable.
What it all comes down to is that research access to APIs is a public-interest issue dressed in technical clothing, and the politics of it determines whether society can understand and hold accountable the systems that increasingly govern it. APIs can be a tremendous accelerant for science and research — democratizing data, enabling collaboration, speeding discovery — and that positive potential is real and worth fighting for. But research access is fragile, dependent on platforms that can revoke it, and frequently treated as expendable, which leaves researchers in a precarious position and society poorer for the studies that don’t get done. And the most politically consequential dimension is the auditing one: the researchers who most need API access are often those studying the platforms and algorithms that most need scrutiny, which gives the powerful a strong incentive to cut off exactly the access that would hold them accountable. The push for stable research access, for SLAs and research programs that protect scholars, for algorithmic transparency through auditable APIs, and for federated and public alternatives that don’t depend on a single platform’s goodwill — all of this is part of the broader political project of ensuring that the systems shaping our lives can be independently studied. Research access to APIs is, in the end, about whether the public can understand the digital systems that govern it, and protecting that access against the platform power that threatens it is one of the most important and most overlooked political causes in the API economy. The science that APIs enable is genuinely valuable; the fight to keep the access that science depends on is genuinely political.
References
- Can API Platforms Ensure Access To Federally Funded Research
- If An Algorithm Impacts Our Life It Should Be Opened Up With An API For Auditing
- Making Scientific Research More Real-Time And Collaborative Using APIs
- Service Level Agreements For Researchers Who Depend On APIs
- The Importance Of APIs In Journalism Right Now
- Having A Program For Researchers Baked Into Your API Operations
- Facebook, Cambridge Analytica, And Knowing What API Consumers Are Doing With Our Data
- The Many Ways In Which APIs Are Taken Away
- Twitter Is The Most Important API
- An Emphasis On The Politics Of APIs
