01
Short invited talks
Concrete cases from app tracking, platform data access, and monetised political content.
3-hour participatory session
What Platforms Hide
How can AI researchers study recommendation, ads, tracking, and monetisation when platforms only show part of the picture?
Concrete cases
App tracking, platform data access, political content monetisation.
Technical problems
Auditing, weak labels, multimodal signals, synthetic data, measurement.
Active room
Bring a method, example, or question.
Session idea
Platforms influence what people encounter every day: which videos are recommended, which products appear first, which ads are delivered, which posts disappear, and which creators get paid.
These are technical systems with public consequences. They shape visibility, money, moderation, and attention, but researchers usually see them from the outside: public posts, links, app traffic, metadata, and limited APIs.
This session is about what computational researchers can do in that situation. We focus on practical methods for studying platforms: auditing, measurement, NLP, multimodal analysis, network analysis, synthetic data, and research designs that work with partial access.
questions
ranking, tracking, ads, monetisation, moderation
methods
audits, measurement, NLP, multimodal ML, networks, synthetic data
Programme
Details to be confirmed
02
Open discussion
Each talk will leave time for questions, examples from the audience, and methodological debate.
03
Small-group frontier exercise
Participants will connect platform problems to methods, disciplines, and possible collaborations.
Participatory core
Platform Research Frontiers
Bring a method. Bring a platform problem. Leave with a sharper question than the one you walked in with.
We will use small groups to turn broad platform concerns into researchable problems: what is worth studying, which methods could help, what breaks in practice, and what could be built next.
01
What should we study?
A viral video, a buried product, a strange ad, a vanished post, a political creator making money.
02
What can each field bring?
Measurement, NLP, multimodal ML, networks, law, social theory, synthetic data, interface design.
03
Where is the frontier?
Labels, ethics, law, incentives, distribution shift, multimodal evidence.
04
What can we do next?
Audits, proxies, simulations, datasets, tools, observatories.
Expected outcome
A short, shared map of platform problems worth studying: what is technically interesting, what society needs to know, and where different fields can push the frontier.
Agenda
Talks and interaction
10 min
Opening and framing
3 x 30 min
Three invited talks, each followed by open discussion
30 min
Small-group frontier exercise
30 min
Group reporting and open discussion
20 min
Networking and potential collaborations
Audience
For AI, ML, NLP, multimodal, network, synthetic data, and computational social science researchers who want their methods near messy public problems. No legal background required.
Confirmed organisers and speakers
Organisers
Thales Bertaglia
Utrecht University
Catalina Goanta
Utrecht University
Konrad Kollnig
Maastricht University & Tech Lab
Adriana
Iamnitchi
Maastricht University
Interested in participating?
Let us know.
If you plan to attend BNAIC and this session sounds relevant, send a quick note. We will use the responses to tune the discussion and group exercise to the people in the room.