Informing and Equipping the Administrators and Moderators of Decentralized Social Media Platforms

Civic Technology Community Group,

In another thread, Rebuilding Local News, we are discussing the vital roles that local news journalism plays in grassroots democracies. We also broached news content markets, attention economics, memes, echo chambers, media bubbles, political polarization, and the mitigation of misinformation and disinformation. We discussed the interdisciplinary topic that negative emotions such as fear and anger can spread, grow, and overwhelm communities as a result of individuals' information gathering, consumption, and sharing behaviors.

I am opening this new thread for brainstorming and discussing new technology ideas towards better informing and equipping the administrators and moderators of decentralized social media platforms.


  1.  How can (deidentified and anonymous) data from decentralized social-media platform instances be used to provide real-time data and services to other server instances' administrators and moderators?
     *   How might server instances' administrators be able to opt into such services?
        *   It could be as easy as entering one or more remote-procedure-call endpoints into a form.
     *   How might end-users be able to opt out?
        *   This might entail a new configuration or setting. Simply, a toggle switch on a settings page.
  2.
Users could be provided with a number of other settings.
     *
Users could opt in to or opt out of receiving warning messages atop displayed news content containing visually graphic content, emotionally charged content, explicit content, and so forth.
        *
As seen on some social media websites, users could be provided with blurred or obscured images atop graphic or explicit content with text warning them, indicating that they can click on the text to bypass the warning.
  3.
Data to be sent to service providers could include users':
     *   anonymous consumption of news content.
     *
anonymous status messages (these can indicate emotional states).
     *
self-reported emotional responses to content items, responses beyond liking<https://www.w3.org/TR/activitypub/#like-activity-outbox> news content or not.
        *
With unfolding advancements to natural-language processing and AI, end-users might be able to click on a response button to enter text into a text field, "this content made me feel _____".
           *
There could be a button next to the text input widget to add a second, third, or fourth such text input field for more complex open-ended reactions.
     *   other, derived indicators of users' emotional states or moods, results of processing content without having to transmit that content.
  4.  What about federated solutions?
     *   As mentioned in 3.d., server instances could process users' data and transmit those results without having to transmit any user data.
     *   Some users might expressly volunteer to share (some or all of) their reactions to news content.
  5.
How can existing and extensible Web standards, e.g., ActivityPub<https://www.w3.org/TR/activitypub/>, be useful for transmitting data between server instances and service providers?
  6.  What about public and scientific access to aggregated data?
     *   Might some or all of these data, including after an amount of delay, be available to the public and the scientific community?
  7.
Real-time natural-language reports, data visualizations, and analytics could be available to administrators and moderators via websites.
     *   Time-critical information, updates, alerts, and alarms could be distributed to administrators or moderators utilizing a variety of communication channels.
  8.
Moderators' tools could be, abstractly, updated in a manner resembling antivirus software updates.
  9.  Moderators' actions and decisions could be collected and processed to disseminate helpful hints to other moderators.
  10. Real-time data could also inform other interoperating systems, e.g., content-distribution algorithms and news recommender systems.

Any thoughts on these ideas? How else can AI, ML, and any other existing and emerging technologies better inform and equip decentralized social media administrators and moderators? Thank you.


Best regards,
Adam