And the next speaker at the Social Media Access Days at the German National Library is Beatrice Cannelli, whose interest is in how national memory institutions’ social media archiving initiatives have been affected by changing data access regimes. Such activities are affected by national legal frameworks, available resources, collection policies and scope, technical limitations, and the Terms of Service of the various platforms.
The latter are justified by user privacy concerns and the protection of sensitive information, but in practice mostly protect the platforms’ own business interests. How these are formulated influences the extent to which content from such platforms is covered by archival initiatives, then; Facebook, for instance, tends to be underrepresented in archives, due to the substantial difficulties inherent in doing so.
Many archival institutions attempting to archive Facebook content were in fact blocked by Facebook at one point or another, and some (like the French National Library) have now indefinitely paused their Facebook collection activities. Some made special requests to Meta to receive special dispensation for archiving, but these were often rebuffed by the platform. The French National Audiovisual Institute’s account for archiving relevant content was even suspended by Meta.
Historically, Twitter had been more open to such archiving by memory institutions; it even had an archive agreement with the US Library of Congress, even if that collaboration never produced any real results. When the Twitter Academic API was launched, many memory institutions applied for access to this service, too, but as non-university institutions largely did not receive approval for API access. At any rate, following the Elon Musk takeover of Twitter, the API was shut down, and current API options offered by Twitter are insufficient or unaffordable for national memory institutions.
At this stage, then, remaining Twitter / X archiving initiatives largely depend on Web crawlers or the archiving of existing Twitter datasets, and there are significant concerns about the long-term archiving of historical Twitter content before it disappears from the platform (or before the platform itself disappears).
Social media archive collections are thus highly vulnerable to platform policy changes that reduce API access or introduce prohibitive pricing levels, or that change platform functionality to make crawler- and scraper-based archiving more difficult. This impacts severely on collection capacity: large-scale collection is made much more difficult, and this especially affects collections around significant time-sensitive events where large datasets need to be captured very rapidly (e.g. before posts disappear).
This in turn impacts on the users of such collections, and on these institutions’ mission to memorialise contemporary public debate for the long term. This also creates a pressing need to include national memory institutions in discussions about research data access, such as those surrounding the EU’s Digital Services Act.
