Cailin O’Connor
@cailinmeister.bsky.social
Philosopher and applied mathematician at UC Irvine. Author of The Misinformation Age and Origins of Unfairness. Irish dancer. Mother. Mother of Chickens.
Invite me next time!
October 30, 2025 at 7:08 PM
Invite me next time!
What can we do? Require COI statements, scrutinize industry research, emphasize independent research when publishing and funding, require data sharing by large platforms, produce and emphasize consensus reports on information tech research
October 24, 2025 at 2:16 PM
What can we do? Require COI statements, scrutinize industry research, emphasize independent research when publishing and funding, require data sharing by large platforms, produce and emphasize consensus reports on information tech research
Science is failing to resist these efforts – there is lax reporting of conflict of interest by funded scientists. And industry sponsored research - with massive data sets and opportunities for interventional research - is routinely published in top journals
October 24, 2025 at 2:15 PM
Science is failing to resist these efforts – there is lax reporting of conflict of interest by funded scientists. And industry sponsored research - with massive data sets and opportunities for interventional research - is routinely published in top journals
And when data is shared, it is to chosen researchers, and often in performative collaborations that make industry look good, while retaining control over the questions asked and methods. There is independence by permission: www.science.org/doi/10.1126/...
Independence by permission
Industry–academy collaboration explores the 2020 US election
www.science.org
October 24, 2025 at 2:15 PM
And when data is shared, it is to chosen researchers, and often in performative collaborations that make industry look good, while retaining control over the questions asked and methods. There is independence by permission: www.science.org/doi/10.1126/...
Information tech is unusual in that companies control the data independent scientists need to study their products – as if Exxon Mobile owned all the thermometers needed to study climate change. By failing to share they can suppress independent research
October 24, 2025 at 2:13 PM
Information tech is unusual in that companies control the data independent scientists need to study their products – as if Exxon Mobile owned all the thermometers needed to study climate change. By failing to share they can suppress independent research
For instance by disproportionately focusing on social media users as the central cause of harms like misinformation and polarization, rather than looking at platform choices related to algorithms or the business model. We call this selective “causal focus”
October 24, 2025 at 2:13 PM
For instance by disproportionately focusing on social media users as the central cause of harms like misinformation and polarization, rather than looking at platform choices related to algorithms or the business model. We call this selective “causal focus”
Notice this means that academics can do good science, exactly how they would otherwise do it, and still play into a research agenda set by industry to accomplish industry goals
October 24, 2025 at 2:12 PM
Notice this means that academics can do good science, exactly how they would otherwise do it, and still play into a research agenda set by industry to accomplish industry goals
H/T to philosophers of science Bennett H and Justin B for describing this kind of “industrial selection” using historical cases from the pharmaceutical industry philpapers.org/rec/HOLEBI
Bennett Holman & Justin Bruner, Experimentation by Industrial Selection - PhilPapers
Industry is a major source of funding for scientific research. There is also a growing concern for how it corrupts researchers faced with conflicts of interest. As such, the debate has ...
philpapers.org
October 24, 2025 at 2:11 PM
H/T to philosophers of science Bennett H and Justin B for describing this kind of “industrial selection” using historical cases from the pharmaceutical industry philpapers.org/rec/HOLEBI
When tech funds academics, they pick who and which projects get funded – and tech disproportionately chooses researchers using methods and aims that tend to yield positive findings for their products
October 24, 2025 at 2:11 PM
When tech funds academics, they pick who and which projects get funded – and tech disproportionately chooses researchers using methods and aims that tend to yield positive findings for their products
Even reasonable internal studies can be design-biased, i.e., make choices that quietly increase chances of positive findings - for example by using stringent tests to assess harms and more lax tests for benefits
October 24, 2025 at 2:08 PM
Even reasonable internal studies can be design-biased, i.e., make choices that quietly increase chances of positive findings - for example by using stringent tests to assess harms and more lax tests for benefits
What are these mechanisms? Tech companies do loads of internal research, some of which they never share or publish. There are basically no tech-sponsored papers finding harms from their products, plenty on benefits
October 24, 2025 at 2:07 PM
What are these mechanisms? Tech companies do loads of internal research, some of which they never share or publish. There are basically no tech-sponsored papers finding harms from their products, plenty on benefits