Today, our information feeds and social media are largely governed by algorithms optimized to maximize engagement, often amplifying the most inflammatory content. With every view, like, and share analyzed to predict and steer our behaviors, we risk becoming subjects of surveillance and manipulation rather than active participants in civic discourse.
In 2025, we will start laying the groundwork for more empathetic and inclusive social networks, with the adoption of what I call “prosocial media.” This is media that doesn’t just capture the attention of users but catalyzes mutual understanding between them. Media that empowers every voice, while fostering the capacity to listen across differences. Media that enables citizens to positively shape the digital public sphere.
One crucial aspect of prosocial media is the ability to allow people to collaboratively add context to potentially misleading information, thereby fostering a more informed discourse. Initiatives like Community Notes on X.com (formerly Twitter) and YouTube, for example, have successfully implemented this for public posts. A recent study, for instance, showed that Twitter Community Notes is an effective tool, reducing the number of retweets of potentially misleading posts by almost half and increasing the probability that a tweet is deleted by the user by 80 percent.
In Taiwan, Cofacts, a community-sourced fact-checking platform, is taking this concept further by empowering citizens to contextualize messages within private groups as well. Launched in 2017 by the civic technology community g0v, the platform was successfully adopted in Thailand in 2019. Research by Cornell University found that Cofacts managed to be quicker and as accurate in dealing with misinformation queries as professional fact-checking sites.
Prosocial media also addresses the centralization of social media platforms and the resulting unhealthy concentration of curation power in the hands of a few tech giants. It does this by using decentralized social networking protocols which enable content to flow seamlessly between different social media platforms. Last year, for instance, Meta’s Threads joined the Fediverse, a group of social media platforms that can communicate with one another, including Mastodon and WordPress. This will eventually allow users on Threads to follow accounts and publish posts on other social networks. In February 2024, another decentralized platform, Bluesky (funded by Twitter founder Jack Dorsey) was also launched to the public.
Decentralization holds the promise of a more democratic internet, where people have greater control over their data and online experiences, leading to a proliferation of local communities, all interconnected through open protocols. This is increasingly valued by users. For instance, research at the University of Cincinnati found that users on decentralized social networks like Mastodon have joined primarily because they could control their information from data mining.
Breaking free of this attention economy will also require bold innovations in the very design of our digital platforms. In 2025, we will start doing that by using AI systems to help us prioritize content that promotes understanding and bridges divides, creating digital spaces that foster genuine dialogue rather than conflict. For instance, Stanford University and Jigsaw, the team created by Google to address global security problems and threats to open societies, have created AI tools that score social media posts and comments based on values like compassion, respect, and curiosity. In April 2024, they published research that demonstrated that ranking posts and comments based on such values significantly reduces reported animosity among users.
In 2025, a new wave of prosocial media platforms will finally start bridging the online divides, highlighting instead the common ground that unites us.
